Skip to content

GitLab

Commit eb93322b authored by mashun1's avatar mashun1
Browse files

dtk24.04.1

parents
Show whitespace changes
Inline Side-by-side
Showing with 4401 additions and 0 deletions
.dockerignore 0 → 100644
View file @ eb93322b
.dockerignore
docker/Dockerfile
.gitignore 0 → 100644
View file @ eb93322b
*.egg*
tryme.ipynb
build/
dist/
test/
temp_output/
temp_fasta/
*pycache*
CONTRIBUTING.md 0 → 100644
View file @ eb93322b
# How to Contribute
We welcome small patches related to bug fixes and documentation, but we do not
plan to make any major changes to this repository.
## Contributor License Agreement
Contributions to this project must be accompanied by a Contributor License
Agreement. You (or your employer) retain the copyright to your contribution,
this simply gives us permission to use and redistribute your contributions as
part of the project. Head over to <https://cla.developers.google.com/> to see
your current agreements on file or to sign a new one.
You generally only need to submit a CLA once, so if you've already submitted one
(even if it was for a different project), you probably don't need to do it
again.
## Code reviews
All submissions, including submissions by project members, require review. We
use GitHub pull requests for this purpose. Consult
[GitHub Help](https://help.github.com/articles/about-pull-requests/) for more
information on using pull requests.
Dockerfile 0 → 100644
View file @ eb93322b
FROM image.sourcefind.cn:5000/dcu/admin/base/jax:0.4.23-ubuntu20.04-dtk24.04.1-py3.10
# RUN apt update
# WORKDIR /app
# WORKDIR /app/softwares
# RUN git clone https://github.com/soedinglab/hh-suite.git
# RUN mkdir -p hh-suite/build && cd hh-suite/build && cmake -DCMAKE_INSTALL_PREFIX=. .. && make -j 4 && make install
# ENV PATH=/app/softwares/hh-suite/build/bin:/app/softwares/hh-suite/build/scripts:$PATH
# WORKDIR /app/softwares
# RUN wget https://github.com/TimoLassmann/kalign/archive/refs/tags/v3.4.0.zip && unzip v3.4.0.zip && cd kalign-3.4.0 && mkdir build && cd build && cmake .. && make && make install
# WORKDIR /app/softwares
# RUN sudo apt install doxygen -y
# RUN wget https://github.com/openmm/openmm/archive/refs/tags/8.0.0.zip && unzip 8.0.0.zip && cd openmm-8.0.0 && mkdir build && cd build && cmake .. && make && sudo make install && sudo make PythonInstall
# WORKDIR /app/softwares
# RUN wget https://github.com/openmm/pdbfixer/archive/refs/tags/1.9.zip && unzip 1.9.zip && cd pdbfixer-1.9 && python setup.py install
# RUN sudo apt install hmmer -y
# WORKDIR /app
# COPY . /app/alphafold2
# RUN ls
# RUN pip install --no-cache-dir -r /app/alphafold2/requirements_dcu.txt -i https://mirrors.ustc.edu.cn/pypi/web/simple
# RUN pip install dm-haiku==0.0.11 flax==0.7.1 jmp==0.0.2 tabulate==0.8.9 --no-deps jax -i https://mirrors.ustc.edu.cn/pypi/web/simple
# RUN pip install orbax==0.1.6 orbax-checkpoint==0.1.6 optax==0.2.2 -i https://mirrors.ustc.edu.cn/pypi/web/simple
# WORKDIR /app/alphafold2
# RUN python setup.py install
LICENSE 0 → 100644
View file @ eb93322b
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
README.md 0 → 100644
View file @ eb93322b
# AF2
## 论文
- [https://www.nature.com/articles/s41586-021-03819-2](https://www.nature.com/articles/s41586-021-03819-2)
## 模型结构
模型核心是一个基于Transformer架构的神经网络，包括两个主要组件：Sequence to Sequence Model和Structure Model，这两个组件通过迭代训练进行优化，以提高其预测准确性。
![img](./docs/alphafold2.png)
## 算法原理
AlphaFold2通过从蛋白质序列和结构数据中提取信息，使用神经网络模型来预测蛋白质三维结构。
![img](./docs/alphafold2_1.png)
## 环境配置
### Docker（方法一）
# 使用该方法不需要下载本仓库，镜像中已包含可运行代码，但需要挂载相应的数据文件
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:alphafold2-dtk24.04.1-py310
docker run --shm-size 100g --network=host --name=alphafold2 --privileged --device=/dev/kfd --device=/dev/dri --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -v 本地数据地址:镜像数据地址 -v /opt/hyhal:/opt/hyhal:ro -it <your IMAGE ID> bash
### Docker（方法二）
docker pull image.sourcefind.cn:5000/dcu/admin/base/jax:0.4.23-ubuntu20.04-dtk24.04.1-py3.10
docker run --shm-size 50g --network=host --name=alphafold2 --privileged --device=/dev/kfd --device=/dev/dri --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -v 项目地址(绝对路径):/home/ -v /opt/hyhal:/opt/hyhal:ro -it <your IMAGE ID> bash
# 1. 一般依赖项安装
pip install -r requirements_dcu.txt
pip install dm-haiku==0.0.11 flax==0.7.1 jmp==0.0.2 tabulate==0.8.9 --no-deps jax
pip install orbax==0.1.6 orbax-checkpoint==0.1.6 optax==0.2.2
python setup.py install
# 2、hh-suite
git clone https://github.com/soedinglab/hh-suite.git
mkdir -p hh-suite/build && cd hh-suite/build
cmake -DCMAKE_INSTALL_PREFIX=. ..
make -j 4 && make install
export PATH="$(pwd)/bin:$(pwd)/scripts:$PATH"
wget https://github.com/TimoLassmann/kalign/archive/refs/tags/v3.4.0.zip
unzip 3.4.0.zip && cd kalign-3.4.0
mkdir build
cd build
cmake ..
make
make test
make install
# 3. openmm + pdbfixer
sudo apt install doxygen
wget https://github.com/openmm/openmm/archive/refs/tags/8.0.0.zip
unzip 8.0.0.zip && cd openmm-8.0.0 && mkdir build && cd build
cmake .. && make && sudo make install && sudo make PythonInstall
wget https://github.com/openmm/pdbfixer/archive/refs/tags/1.9.zip
unzip 1.9.zip && cd pdbfixer-1.9 && python setup.py install
## 数据集
推荐使用AlphaFold2中的开源数据集，包括BFD、MGnify、PDB70、Uniclust、Uniref90等,数据集大小约2.62TB。数据集格式如下：
```
$DOWNLOAD_DIR/
bfd/
bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt_hhm.ffindex
bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt_hhm.ffdata
bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt_cs219.ffindex
...
mgnify/
mgy_clusters_2022_05.fa
params/
params_model_1.npz
params_model_2.npz
params_model_3.npz
...
pdb70/
pdb_filter.dat
pdb70_hhm.ffindex
pdb70_hhm.ffdata
...
pdb_mmcif/
mmcif_files/
100d.cif
101d.cif
101m.cif
...
obsolete.dat
pdb_seqres/
pdb_seqres.txt
small_bfd/
bfd-first_non_consensus_sequences.fasta
uniref30/
UniRef30_2021_03_hhm.ffindex
UniRef30_2021_03_hhm.ffdata
UniRef30_2021_03_cs219.ffindex
...
uniprot/
uniprot.fasta
uniref90/
uniref90.fasta
```
此处提供了一个脚本download_all_data.sh用于下载使用的数据集和模型文件：
./scripts/download_all_data.sh 数据集下载目录
数据集快速下载中心：[SCNet AIDatasets](http://113.200.138.88:18080/aidatasets) ，项目中数据集可从快速下载通道下载：[alphafold](http://113.200.138.88:18080/aidatasets/project-dependency/alphafold) 。
## 推理
分别提供了基于Jax的单体和多体的推理脚本.
```bash
# 进入工程目录
cd alphafold2_jax
```
### 单体
```bash
./run_monomer.sh
```
单体推理参数说明：download_dir为数据集下载目录，monomer.fasta为推理的单体序列；`--output_dir`为输出目录；`model_names`为推理的模型名称，`--model_preset=monomer`为单体模型配置；`--run_relax=true`为进行relax操作；`--use_gpu_relax=true`为使用gpu进行relax操作（速度更快，但可能不太稳定），`--use_gpu_relax=false`为使用CPU进行relax操作（速度慢，但稳定）；若添加--use_precomputed_msas=true则可以加载已有的MSAs，否则默认运行MSA工具。
### 多体
```bash
./run_multimer.sh
```
多体推理参数说明：multimer.fasta为推理的多体序列，`--model_preset=multimer`为多体模型配置；`--num_multimer_predictions_per_model`为每个模型预测数量，其他参数同单体推理参数说明一致。
## result
`--output_dir`目录结构如下：
```
<target_name>/
features.pkl
ranked_{0,1,2,3,4}.pdb
ranking_debug.json
relaxed_model_{1,2,3,4,5}.pdb
result_model_{1,2,3,4,5}.pkl
timings.json
unrelaxed_model_{1,2,3,4,5}.pdb
msas/
bfd_uniclust_hits.a3m
mgnify_hits.sto
uniref90_hits.sto
...
```
[查看蛋白质3D结构](https://www.pdbus.org/3d-view)
ID: 8U23
蓝色的为预测结构，黄色为真实结构
![alt text](image.png)
### 精度
无
## 应用场景
### 算法类别
蛋白质预测
### 热点应用行业
医疗,科研,教育
## 预训练权重
预训练权重快速下载中心：[SCNet AIModels](http://113.200.138.88:18080/aimodels) ，项目中的预训练权重可从快速下载通道下载：[alphafold](http://113.200.138.88:18080/aimodels/findsource-dependency/alphafold-params) 。
## 源码仓库及问题反馈
* [https://developer.hpccube.com/codes/modelzoo/alphafold2_jax](https://developer.hpccube.com/codes/modelzoo/alphafold2_jax)
## 参考
* [https://github.com/deepmind/alphafold](https://github.com/deepmind/alphafold)
README_official.md 0 → 100644
View file @ eb93322b
![header](imgs/header.jpg)
# AlphaFold
This package provides an implementation of the inference pipeline of AlphaFold
v2. For simplicity, we refer to this model as AlphaFold throughout the rest of
this document.
We also provide:
1. An implementation of AlphaFold-Multimer. This represents a work in progress
and AlphaFold-Multimer isn't expected to be as stable as our monomer
AlphaFold system. [Read the guide](#updating-existing-installation) for how
to upgrade and update code.
2. The [technical note](docs/technical_note_v2.3.0.md) containing the models
and inference procedure for an updated AlphaFold v2.3.0.
3. A [CASP15 baseline](docs/casp15_predictions.zip) set of predictions along
with documentation of any manual interventions performed.
Any publication that discloses findings arising from using this source code or
the model parameters should [cite](#citing-this-work) the
[AlphaFold paper](https://doi.org/10.1038/s41586-021-03819-2) and, if
applicable, the
[AlphaFold-Multimer paper](https://www.biorxiv.org/content/10.1101/2021.10.04.463034v1).
Please also refer to the
[Supplementary Information](https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-021-03819-2/MediaObjects/41586_2021_3819_MOESM1_ESM.pdf)
for a detailed description of the method.
**You can use a slightly simplified version of AlphaFold with
[this Colab notebook](https://colab.research.google.com/github/deepmind/alphafold/blob/main/notebooks/AlphaFold.ipynb)**
or community-supported versions (see below).
If you have any questions, please contact the AlphaFold team at
[alphafold@deepmind.com](mailto:alphafold@deepmind.com).
![CASP14 predictions](imgs/casp14_predictions.gif)
## Installation and running your first prediction
You will need a machine running Linux, AlphaFold does not support other
operating systems. Full installation requires up to 3 TB of disk space to keep
genetic databases (SSD storage is recommended) and a modern NVIDIA GPU (GPUs
with more memory can predict larger protein structures).
Please follow these steps:
1. Install [Docker](https://www.docker.com/).
* Install
[NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html)
for GPU support.
* Setup running
[Docker as a non-root user](https://docs.docker.com/engine/install/linux-postinstall/#manage-docker-as-a-non-root-user).
1. Clone this repository and `cd` into it.
```bash
git clone https://github.com/deepmind/alphafold.git
cd ./alphafold
```
1. Download genetic databases and model parameters:
* Install `aria2c`. On most Linux distributions it is available via the
package manager as the `aria2` package (on Debian-based distributions this
can be installed by running `sudo apt install aria2`).
* Please use the script `scripts/download_all_data.sh` to download
and set up full databases. This may take substantial time (download size is
556 GB), so we recommend running this script in the background:
```bash
scripts/download_all_data.sh <DOWNLOAD_DIR> > download.log 2> download_all.log &
```
* **Note: The download directory `<DOWNLOAD_DIR>` should *not* be a
subdirectory in the AlphaFold repository directory.** If it is, the Docker
build will be slow as the large databases will be copied into the docker
build context.
* It is possible to run AlphaFold with reduced databases; please refer to
the [complete documentation](#genetic-databases).
1. Check that AlphaFold will be able to use a GPU by running:
```bash
docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi
```
The output of this command should show a list of your GPUs. If it doesn't,
check if you followed all steps correctly when setting up the
[NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html)
or take a look at the following
[NVIDIA Docker issue](https://github.com/NVIDIA/nvidia-docker/issues/1447#issuecomment-801479573).
If you wish to run AlphaFold using Singularity (a common containerization
platform on HPC systems) we recommend using some of the third party Singularity
setups as linked in https://github.com/deepmind/alphafold/issues/10 or
https://github.com/deepmind/alphafold/issues/24.
1. Build the Docker image:
```bash
docker build -f docker/Dockerfile -t alphafold .
```
If you encounter the following error:
```
W: GPG error: https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64 InRelease: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY A4B469963BF863CC
E: The repository 'https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64 InRelease' is not signed.
```
use the workaround described in
https://github.com/deepmind/alphafold/issues/463#issuecomment-1124881779.
1. Install the `run_docker.py` dependencies. Note: You may optionally wish to
create a
[Python Virtual Environment](https://docs.python.org/3/tutorial/venv.html)
to prevent conflicts with your system's Python environment.
```bash
pip3 install -r docker/requirements.txt
```
1. Make sure that the output directory exists (the default is `/tmp/alphafold`)
and that you have sufficient permissions to write into it.
1. Run `run_docker.py` pointing to a FASTA file containing the protein
sequence(s) for which you wish to predict the structure (`--fasta_paths`
parameter). AlphaFold will search for the available templates before the
date specified by the `--max_template_date` parameter; this could be used to
avoid certain templates during modeling. `--data_dir` is the directory with
downloaded genetic databases and `--output_dir` is the absolute path to the
output directory.
```bash
python3 docker/run_docker.py \
--fasta_paths=your_protein.fasta \
--max_template_date=2022-01-01 \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
1. Once the run is over, the output directory shall contain predicted
structures of the target protein. Please check the documentation below for
additional options and troubleshooting tips.
### Genetic databases
This step requires `aria2c` to be installed on your machine.
AlphaFold needs multiple genetic (sequence) databases to run:
* [BFD](https://bfd.mmseqs.com/),
* [MGnify](https://www.ebi.ac.uk/metagenomics/),
* [PDB70](http://wwwuser.gwdg.de/~compbiol/data/hhsuite/databases/hhsuite_dbs/),
* [PDB](https://www.rcsb.org/) (structures in the mmCIF format),
* [PDB seqres](https://www.rcsb.org/) – only for AlphaFold-Multimer,
* [UniRef30 (FKA UniClust30)](https://uniclust.mmseqs.com/),
* [UniProt](https://www.uniprot.org/uniprot/) – only for AlphaFold-Multimer,
* [UniRef90](https://www.uniprot.org/help/uniref).
We provide a script `scripts/download_all_data.sh` that can be used to download
and set up all of these databases:
* Recommended default:
```bash
scripts/download_all_data.sh <DOWNLOAD_DIR>
```
will download the full databases.
* With `reduced_dbs` parameter:
```bash
scripts/download_all_data.sh <DOWNLOAD_DIR> reduced_dbs
```
will download a reduced version of the databases to be used with the
`reduced_dbs` database preset. This shall be used with the corresponding
AlphaFold parameter `--db_preset=reduced_dbs` later during the AlphaFold run
(please see [AlphaFold parameters](#running-alphafold) section).
:ledger: **Note: The download directory `<DOWNLOAD_DIR>` should *not* be a
subdirectory in the AlphaFold repository directory.** If it is, the Docker build
will be slow as the large databases will be copied during the image creation.
We don't provide exactly the database versions used in CASP14 – see the
[note on reproducibility](#note-on-casp14-reproducibility). Some of the
databases are mirrored for speed, see [mirrored databases](#mirrored-databases).
:ledger: **Note: The total download size for the full databases is around 556 GB
and the total size when unzipped is 2.62 TB. Please make sure you have a large
enough hard drive space, bandwidth and time to download. We recommend using an
SSD for better genetic search performance.**
:ledger: **Note: If the download directory and datasets don't have full read and
write permissions, it can cause errors with the MSA tools, with opaque
(external) error messages. Please ensure the required permissions are applied,
e.g. with the `sudo chmod 755 --recursive "$DOWNLOAD_DIR"` command.**
The `download_all_data.sh` script will also download the model parameter files.
Once the script has finished, you should have the following directory structure:
```
$DOWNLOAD_DIR/ # Total: ~ 2.62 TB (download: 556 GB)
bfd/ # ~ 1.8 TB (download: 271.6 GB)
# 6 files.
mgnify/ # ~ 120 GB (download: 67 GB)
mgy_clusters_2022_05.fa
params/ # ~ 5.3 GB (download: 5.3 GB)
# 5 CASP14 models,
# 5 pTM models,
# 5 AlphaFold-Multimer models,
# LICENSE,
# = 16 files.
pdb70/ # ~ 56 GB (download: 19.5 GB)
# 9 files.
pdb_mmcif/ # ~ 238 GB (download: 43 GB)
mmcif_files/
# About 199,000 .cif files.
obsolete.dat
pdb_seqres/ # ~ 0.2 GB (download: 0.2 GB)
pdb_seqres.txt
small_bfd/ # ~ 17 GB (download: 9.6 GB)
bfd-first_non_consensus_sequences.fasta
uniref30/ # ~ 206 GB (download: 52.5 GB)
# 7 files.
uniprot/ # ~ 105 GB (download: 53 GB)
uniprot.fasta
uniref90/ # ~ 67 GB (download: 34 GB)
uniref90.fasta
```
`bfd/` is only downloaded if you download the full databases, and `small_bfd/`
is only downloaded if you download the reduced databases.
### Model parameters
While the AlphaFold code is licensed under the Apache 2.0 License, the AlphaFold
parameters and CASP15 prediction data are made available under the terms of the
CC BY 4.0 license. Please see the [Disclaimer](#license-and-disclaimer) below
for more detail.
The AlphaFold parameters are available from
https://storage.googleapis.com/alphafold/alphafold_params_2022-12-06.tar, and
are downloaded as part of the `scripts/download_all_data.sh` script. This script
will download parameters for:
* 5 models which were used during CASP14, and were extensively validated for
structure prediction quality (see Jumper et al. 2021, Suppl. Methods 1.12
for details).
* 5 pTM models, which were fine-tuned to produce pTM (predicted TM-score) and
(PAE) predicted aligned error values alongside their structure predictions
(see Jumper et al. 2021, Suppl. Methods 1.9.7 for details).
* 5 AlphaFold-Multimer models that produce pTM and PAE values alongside their
structure predictions.
### Updating existing installation
If you have a previous version you can either reinstall fully from scratch
(remove everything and run the setup from scratch) or you can do an incremental
update that will be significantly faster but will require a bit more work. Make
sure you follow these steps in the exact order they are listed below:
1. **Update the code.**
* Go to the directory with the cloned AlphaFold repository and run `git
fetch origin main` to get all code updates.
1. **Update the UniProt, UniRef, MGnify and PDB seqres databases.**
* Remove `<DOWNLOAD_DIR>/uniprot`.
* Run `scripts/download_uniprot.sh <DOWNLOAD_DIR>`.
* Remove `<DOWNLOAD_DIR>/uniclust30`.
* Run `scripts/download_uniref30.sh <DOWNLOAD_DIR>`.
* Remove `<DOWNLOAD_DIR>/uniref90`.
* Run `scripts/download_uniref90.sh <DOWNLOAD_DIR>`.
* Remove `<DOWNLOAD_DIR>/mgnify`.
* Run `scripts/download_mgnify.sh <DOWNLOAD_DIR>`.
* Remove `<DOWNLOAD_DIR>/pdb_mmcif`. It is needed to have PDB SeqRes and
PDB from exactly the same date. Failure to do this step will result in
potential errors when searching for templates when running
AlphaFold-Multimer.
* Run `scripts/download_pdb_mmcif.sh <DOWNLOAD_DIR>`.
* Run `scripts/download_pdb_seqres.sh <DOWNLOAD_DIR>`.
1. **Update the model parameters.**
* Remove the old model parameters in `<DOWNLOAD_DIR>/params`.
* Download new model parameters using
`scripts/download_alphafold_params.sh <DOWNLOAD_DIR>`.
1. **Follow [Running AlphaFold](#running-alphafold).**
#### Using deprecated model weights
To use the deprecated v2.2.0 AlphaFold-Multimer model weights:
1. Change `SOURCE_URL` in `scripts/download_alphafold_params.sh` to
`https://storage.googleapis.com/alphafold/alphafold_params_2022-03-02.tar`,
and download the old parameters.
2. Change the `_v3` to `_v2` in the multimer `MODEL_PRESETS` in `config.py`.
To use the deprecated v2.1.0 AlphaFold-Multimer model weights:
1. Change `SOURCE_URL` in `scripts/download_alphafold_params.sh` to
`https://storage.googleapis.com/alphafold/alphafold_params_2022-01-19.tar`,
and download the old parameters.
2. Remove the `_v3` in the multimer `MODEL_PRESETS` in `config.py`.
## Running AlphaFold
**The simplest way to run AlphaFold is using the provided Docker script.** This
was tested on Google Cloud with a machine using the `nvidia-gpu-cloud-image`
with 12 vCPUs, 85 GB of RAM, a 100 GB boot disk, the databases on an additional
3 TB disk, and an A100 GPU. For your first run, please follow the instructions
from [Installation and running your first prediction](#installation-and-running-your-first-prediction)
section.
1. By default, Alphafold will attempt to use all visible GPU devices. To use a
subset, specify a comma-separated list of GPU UUID(s) or index(es) using the
`--gpu_devices` flag. See
[GPU enumeration](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/user-guide.html#gpu-enumeration)
for more details.
1. You can control which AlphaFold model to run by adding the `--model_preset=`
flag. We provide the following models:
* **monomer**: This is the original model used at CASP14 with no
ensembling.
* **monomer\_casp14**: This is the original model used at CASP14 with
`num_ensemble=8`, matching our CASP14 configuration. This is largely
provided for reproducibility as it is 8x more computationally expensive
for limited accuracy gain (+0.1 average GDT gain on CASP14 domains).
* **monomer\_ptm**: This is the original CASP14 model fine tuned with the
pTM head, providing a pairwise confidence measure. It is slightly less
accurate than the normal monomer model.
* **multimer**: This is the [AlphaFold-Multimer](#citing-this-work) model.
To use this model, provide a multi-sequence FASTA file. In addition, the
UniProt database should have been downloaded.
1. You can control MSA speed/quality tradeoff by adding
`--db_preset=reduced_dbs` or `--db_preset=full_dbs` to the run command. We
provide the following presets:
* **reduced\_dbs**: This preset is optimized for speed and lower hardware
requirements. It runs with a reduced version of the BFD database. It
requires 8 CPU cores (vCPUs), 8 GB of RAM, and 600 GB of disk space.
* **full\_dbs**: This runs with all genetic databases used at CASP14.
Running the command above with the `monomer` model preset and the
`reduced_dbs` data preset would look like this:
```bash
python3 docker/run_docker.py \
--fasta_paths=T1050.fasta \
--max_template_date=2020-05-14 \
--model_preset=monomer \
--db_preset=reduced_dbs \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
1. After generating the predicted model, AlphaFold runs a relaxation
step to improve local geometry. By default, only the best model (by
pLDDT) is relaxed (`--models_to_relax=best`), but also all of the models
(`--models_to_relax=all`) or none of the models (`--models_to_relax=none`)
can be relaxed.
1. The relaxation step can be run on GPU (faster, but could be less stable) or
CPU (slow, but stable). This can be controlled with `--enable_gpu_relax=true`
(default) or `--enable_gpu_relax=false`.
1. AlphaFold can re-use MSAs (multiple sequence alignments) for the same
sequence via `--use_precomputed_msas=true` option; this can be useful for
trying different AlphaFold parameters. This option assumes that the
directory structure generated by the first AlphaFold run in the output
directory exists and that the protein sequence is the same.
### Running AlphaFold-Multimer
All steps are the same as when running the monomer system, but you will have to
* provide an input fasta with multiple sequences,
* set `--model_preset=multimer`,
An example that folds a protein complex `multimer.fasta`:
```bash
python3 docker/run_docker.py \
--fasta_paths=multimer.fasta \
--max_template_date=2020-05-14 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
By default the multimer system will run 5 seeds per model (25 total predictions)
for a small drop in accuracy you may wish to run a single seed per model. This
can be done via the `--num_multimer_predictions_per_model` flag, e.g. set it to
`--num_multimer_predictions_per_model=1` to run a single seed per model.
### AlphaFold prediction speed
The table below reports prediction runtimes for proteins of various lengths. We
only measure unrelaxed structure prediction with three recycles while
excluding runtimes from MSA and template search. When running
`docker/run_docker.py` with `--benchmark=true`, this runtime is stored in
`timings.json`. All runtimes are from a single A100 NVIDIA GPU. Prediction
speed on A100 for smaller structures can be improved by increasing
`global_config.subbatch_size` in `alphafold/model/config.py`.
No. residues | Prediction time (s)
-----------: | ------------------:
100 | 4.9
200 | 7.7
300 | 13
400 | 18
500 | 29
600 | 36
700 | 53
800 | 60
900 | 91
1,000 | 96
1,100 | 140
1,500 | 280
2,000 | 450
2,500 | 969
3,000 | 1,240
3,500 | 2,465
4,000 | 5,660
4,500 | 12,475
5,000 | 18,824
### Examples
Below are examples on how to use AlphaFold in different scenarios.
#### Folding a monomer
Say we have a monomer with the sequence `<SEQUENCE>`. The input fasta should be:
```fasta
>sequence_name
<SEQUENCE>
```
Then run the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=monomer.fasta \
--max_template_date=2021-11-01 \
--model_preset=monomer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
#### Folding a homomer
Say we have a homomer with 3 copies of the same sequence `<SEQUENCE>`. The input
fasta should be:
```fasta
>sequence_1
<SEQUENCE>
>sequence_2
<SEQUENCE>
>sequence_3
<SEQUENCE>
```
Then run the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=homomer.fasta \
--max_template_date=2021-11-01 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
#### Folding a heteromer
Say we have an A2B3 heteromer, i.e. with 2 copies of `<SEQUENCE A>` and 3 copies
of `<SEQUENCE B>`. The input fasta should be:
```fasta
>sequence_1
<SEQUENCE A>
>sequence_2
<SEQUENCE A>
>sequence_3
<SEQUENCE B>
>sequence_4
<SEQUENCE B>
>sequence_5
<SEQUENCE B>
```
Then run the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=heteromer.fasta \
--max_template_date=2021-11-01 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
#### Folding multiple monomers one after another
Say we have a two monomers, `monomer1.fasta` and `monomer2.fasta`.
We can fold both sequentially by using the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=monomer1.fasta,monomer2.fasta \
--max_template_date=2021-11-01 \
--model_preset=monomer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
#### Folding multiple multimers one after another
Say we have a two multimers, `multimer1.fasta` and `multimer2.fasta`.
We can fold both sequentially by using the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=multimer1.fasta,multimer2.fasta \
--max_template_date=2021-11-01 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
```
### AlphaFold output
The outputs will be saved in a subdirectory of the directory provided via the
`--output_dir` flag of `run_docker.py` (defaults to `/tmp/alphafold/`). The
outputs include the computed MSAs, unrelaxed structures, relaxed structures,
ranked structures, raw model outputs, prediction metadata, and section timings.
The `--output_dir` directory will have the following structure:
```
<target_name>/
features.pkl
ranked_{0,1,2,3,4}.pdb
ranking_debug.json
relax_metrics.json
relaxed_model_{1,2,3,4,5}.pdb
result_model_{1,2,3,4,5}.pkl
timings.json
unrelaxed_model_{1,2,3,4,5}.pdb
msas/
bfd_uniref_hits.a3m
mgnify_hits.sto
uniref90_hits.sto
```
The contents of each output file are as follows:
* `features.pkl` – A `pickle` file containing the input feature NumPy arrays
used by the models to produce the structures.
* `unrelaxed_model_*.pdb` – A PDB format text file containing the predicted
structure, exactly as outputted by the model.
* `relaxed_model_*.pdb` – A PDB format text file containing the predicted
structure, after performing an Amber relaxation procedure on the unrelaxed
structure prediction (see Jumper et al. 2021, Suppl. Methods 1.8.6 for
details).
* `ranked_*.pdb` – A PDB format text file containing the predicted structures,
after reordering by model confidence. Here `ranked_i.pdb` should contain
the prediction with the (`i + 1`)-th highest confidence (so that
`ranked_0.pdb` has the highest confidence). To rank model confidence, we use
predicted LDDT (pLDDT) scores (see Jumper et al. 2021, Suppl. Methods 1.9.6
for details). If `--models_to_relax=all` then all ranked structures are
relaxed. If `--models_to_relax=best` then only `ranked_0.pdb` is relaxed
(the rest are unrelaxed). If `--models_to_relax=none`, then the ranked
structures are all unrelaxed.
* `ranking_debug.json` – A JSON format text file containing the pLDDT values
used to perform the model ranking, and a mapping back to the original model
names.
* `relax_metrics.json` – A JSON format text file containing relax metrics, for
instance remaining violations.
* `timings.json` – A JSON format text file containing the times taken to run
each section of the AlphaFold pipeline.
* `msas/` - A directory containing the files describing the various genetic
tool hits that were used to construct the input MSA.
* `result_model_*.pkl` – A `pickle` file containing a nested dictionary of the
various NumPy arrays directly produced by the model. In addition to the
output of the structure module, this includes auxiliary outputs such as:
* Distograms (`distogram/logits` contains a NumPy array of shape [N_res,
N_res, N_bins] and `distogram/bin_edges` contains the definition of the
bins).
* Per-residue pLDDT scores (`plddt` contains a NumPy array of shape
[N_res] with the range of possible values from `0` to `100`, where `100`
means most confident). This can serve to identify sequence regions
predicted with high confidence or as an overall per-target confidence
score when averaged across residues.
* Present only if using pTM models: predicted TM-score (`ptm` field
contains a scalar). As a predictor of a global superposition metric,
this score is designed to also assess whether the model is confident in
the overall domain packing.
* Present only if using pTM models: predicted pairwise aligned errors
(`predicted_aligned_error` contains a NumPy array of shape [N_res,
N_res] with the range of possible values from `0` to
`max_predicted_aligned_error`, where `0` means most confident). This can
serve for a visualisation of domain packing confidence within the
structure.
The pLDDT confidence measure is stored in the B-factor field of the output PDB
files (although unlike a B-factor, higher pLDDT is better, so care must be taken
when using for tasks such as molecular replacement).
This code has been tested to match mean top-1 accuracy on a CASP14 test set with
pLDDT ranking over 5 model predictions (some CASP targets were run with earlier
versions of AlphaFold and some had manual interventions; see our forthcoming
publication for details). Some targets such as T1064 may also have high
individual run variance over random seeds.
## Inferencing many proteins
The provided inference script is optimized for predicting the structure of a
single protein, and it will compile the neural network to be specialized to
exactly the size of the sequence, MSA, and templates. For large proteins, the
compile time is a negligible fraction of the runtime, but it may become more
significant for small proteins or if the multi-sequence alignments are already
precomputed. In the bulk inference case, it may make sense to use our
`make_fixed_size` function to pad the inputs to a uniform size, thereby reducing
the number of compilations required.
We do not provide a bulk inference script, but it should be straightforward to
develop on top of the `RunModel.predict` method with a parallel system for
precomputing multi-sequence alignments. Alternatively, this script can be run
repeatedly with only moderate overhead.
## Note on CASP14 reproducibility
AlphaFold's output for a small number of proteins has high inter-run variance,
and may be affected by changes in the input data. The CASP14 target T1064 is a
notable example; the large number of SARS-CoV-2-related sequences recently
deposited changes its MSA significantly. This variability is somewhat mitigated
by the model selection process; running 5 models and taking the most confident.
To reproduce the results of our CASP14 system as closely as possible you must
use the same database versions we used in CASP. These may not match the default
versions downloaded by our scripts.
For genetics:
* UniRef90:
[v2020_01](https://ftp.uniprot.org/pub/databases/uniprot/previous_releases/release-2020_01/uniref/)
* MGnify:
[v2018_12](http://ftp.ebi.ac.uk/pub/databases/metagenomics/peptide_database/2018_12/)
* Uniclust30: [v2018_08](http://wwwuser.gwdg.de/~compbiol/uniclust/2018_08/)
* BFD: [only version available](https://bfd.mmseqs.com/)
For templates:
* PDB: (downloaded 2020-05-14)
* PDB70:
[2020-05-13](http://wwwuser.gwdg.de/~compbiol/data/hhsuite/databases/hhsuite_dbs/old-releases/pdb70_from_mmcif_200513.tar.gz)
An alternative for templates is to use the latest PDB and PDB70, but pass the
flag `--max_template_date=2020-05-14`, which restricts templates only to
structures that were available at the start of CASP14.
## Citing this work
If you use the code or data in this package, please cite:
```bibtex
@Article{AlphaFold2021,
author = {Jumper, John and Evans, Richard and Pritzel, Alexander and Green, Tim and Figurnov, Michael and Ronneberger, Olaf and Tunyasuvunakool, Kathryn and Bates, Russ and {\v{Z}}{\'\i}dek, Augustin and Potapenko, Anna and Bridgland, Alex and Meyer, Clemens and Kohl, Simon A A and Ballard, Andrew J and Cowie, Andrew and Romera-Paredes, Bernardino and Nikolov, Stanislav and Jain, Rishub and Adler, Jonas and Back, Trevor and Petersen, Stig and Reiman, David and Clancy, Ellen and Zielinski, Michal and Steinegger, Martin and Pacholska, Michalina and Berghammer, Tamas and Bodenstein, Sebastian and Silver, David and Vinyals, Oriol and Senior, Andrew W and Kavukcuoglu, Koray and Kohli, Pushmeet and Hassabis, Demis},
journal = {Nature},
title = {Highly accurate protein structure prediction with {AlphaFold}},
year = {2021},
volume = {596},
number = {7873},
pages = {583--589},
doi = {10.1038/s41586-021-03819-2}
}
```
In addition, if you use the AlphaFold-Multimer mode, please cite:
```bibtex
@article {AlphaFold-Multimer2021,
author = {Evans, Richard and O{\textquoteright}Neill, Michael and Pritzel, Alexander and Antropova, Natasha and Senior, Andrew and Green, Tim and {\v{Z}}{\'\i}dek, Augustin and Bates, Russ and Blackwell, Sam and Yim, Jason and Ronneberger, Olaf and Bodenstein, Sebastian and Zielinski, Michal and Bridgland, Alex and Potapenko, Anna and Cowie, Andrew and Tunyasuvunakool, Kathryn and Jain, Rishub and Clancy, Ellen and Kohli, Pushmeet and Jumper, John and Hassabis, Demis},
journal = {bioRxiv},
title = {Protein complex prediction with AlphaFold-Multimer},
year = {2021},
elocation-id = {2021.10.04.463034},
doi = {10.1101/2021.10.04.463034},
URL = {https://www.biorxiv.org/content/early/2021/10/04/2021.10.04.463034},
eprint = {https://www.biorxiv.org/content/early/2021/10/04/2021.10.04.463034.full.pdf},
}
```
## Community contributions
Colab notebooks provided by the community (please note that these notebooks may
vary from our full AlphaFold system and we did not validate their accuracy):
* The
[ColabFold AlphaFold2 notebook](https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb)
by Martin Steinegger, Sergey Ovchinnikov and Milot Mirdita, which uses an
API hosted at the Södinglab based on the MMseqs2 server
[(Mirdita et al. 2019, Bioinformatics)](https://academic.oup.com/bioinformatics/article/35/16/2856/5280135)
for the multiple sequence alignment creation.
## Acknowledgements
AlphaFold communicates with and/or references the following separate libraries
and packages:
* [Abseil](https://github.com/abseil/abseil-py)
* [Biopython](https://biopython.org)
* [Chex](https://github.com/deepmind/chex)
* [Colab](https://research.google.com/colaboratory/)
* [Docker](https://www.docker.com)
* [HH Suite](https://github.com/soedinglab/hh-suite)
* [HMMER Suite](http://eddylab.org/software/hmmer)
* [Haiku](https://github.com/deepmind/dm-haiku)
* [Immutabledict](https://github.com/corenting/immutabledict)
* [JAX](https://github.com/google/jax/)
* [Kalign](https://msa.sbc.su.se/cgi-bin/msa.cgi)
* [matplotlib](https://matplotlib.org/)
* [ML Collections](https://github.com/google/ml_collections)
* [NumPy](https://numpy.org)
* [OpenMM](https://github.com/openmm/openmm)
* [OpenStructure](https://openstructure.org)
* [pandas](https://pandas.pydata.org/)
* [pymol3d](https://github.com/avirshup/py3dmol)
* [SciPy](https://scipy.org)
* [Sonnet](https://github.com/deepmind/sonnet)
* [TensorFlow](https://github.com/tensorflow/tensorflow)
* [Tree](https://github.com/deepmind/tree)
* [tqdm](https://github.com/tqdm/tqdm)
We thank all their contributors and maintainers!
## Get in Touch
If you have any questions not covered in this overview, please contact the
AlphaFold team at [alphafold@deepmind.com](mailto:alphafold@deepmind.com).
We would love to hear your feedback and understand how AlphaFold has been useful
in your research. Share your stories with us at
[alphafold@deepmind.com](mailto:alphafold@deepmind.com).
## License and Disclaimer
This is not an officially supported Google product.
Copyright 2022 DeepMind Technologies Limited.
### AlphaFold Code License
Licensed under the Apache License, Version 2.0 (the "License"); you may not use
this file except in compliance with the License. You may obtain a copy of the
License at https://www.apache.org/licenses/LICENSE-2.0.
Unless required by applicable law or agreed to in writing, software distributed
under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
### Model Parameters License
The AlphaFold parameters are made available under the terms of the Creative
Commons Attribution 4.0 International (CC BY 4.0) license. You can find details
at: https://creativecommons.org/licenses/by/4.0/legalcode
### Third-party software
Use of the third-party software, libraries or code referred to in the
[Acknowledgements](#acknowledgements) section above may be governed by separate
terms and conditions or license provisions. Your use of the third-party
software, libraries or code is subject to any such terms and you should check
that you can comply with any applicable restrictions or terms and conditions
before use.
### Mirrored Databases
The following databases have been mirrored by DeepMind, and are available with
reference to the following:
* [BFD](https://bfd.mmseqs.com/) (unmodified), by Steinegger M. and Söding J.,
available under a
[Creative Commons Attribution-ShareAlike 4.0 International License](http://creativecommons.org/licenses/by-sa/4.0/).
* [BFD](https://bfd.mmseqs.com/) (modified), by Steinegger M. and Söding J.,
modified by DeepMind, available under a
[Creative Commons Attribution-ShareAlike 4.0 International License](http://creativecommons.org/licenses/by-sa/4.0/).
See the Methods section of the
[AlphaFold proteome paper](https://www.nature.com/articles/s41586-021-03828-1)
for details.
* [Uniref30: v2021_03](http://wwwuser.gwdg.de/~compbiol/uniclust/2021_03/)
(unmodified), by Mirdita M. et al., available under a
[Creative Commons Attribution-ShareAlike 4.0 International License](http://creativecommons.org/licenses/by-sa/4.0/).
* [MGnify: v2022_05](http://ftp.ebi.ac.uk/pub/databases/metagenomics/peptide_database/2022_05/README.txt)
(unmodified), by Mitchell AL et al., available free of all copyright
restrictions and made fully and freely available for both non-commercial and
commercial use under
[CC0 1.0 Universal (CC0 1.0) Public Domain Dedication](https://creativecommons.org/publicdomain/zero/1.0/).
afdb/README.md 0 → 100644
View file @ eb93322b
# AlphaFold Protein Structure Database
## Introduction
The AlphaFold UniProt release (214M predictions) is hosted on
[Google Cloud Public Datasets](https://console.cloud.google.com/marketplace/product/bigquery-public-data/deepmind-alphafold),
and is available to download at no cost under a
[CC-BY-4.0 licence](http://creativecommons.org/licenses/by/4.0/legalcode). The
dataset is in a Cloud Storage bucket, and metadata is available on BigQuery. A
Google Cloud account is required for the download, but the data can be freely
used under the terms of the
[CC-BY 4.0 Licence](http://creativecommons.org/licenses/by/4.0/legalcode).
This document provides an overview of how to access and download the dataset for
different use cases. Please refer to the [AlphaFold database FAQ](https://www.alphafold.com/faq)
for further information on what proteins are in the database and a changelog of
releases.
:ledger: **Note: The full dataset is difficult to manipulate without significant
computational resources (the size of the dataset is 23 TiB, 3 * 214M files).**
There are also alternatives to downloading the full dataset:
1. Download a premade subset (covering important species / Swiss-Prot) via our
[download page](https://alphafold.ebi.ac.uk/download).
2. Download a custom subset of the data. See below.
If you need to download the full dataset then please see the "Bulk download"
section. See "Creating a Google Cloud Account" below for more information on how
to avoid any surprise costs when using Google Cloud Public Datasets.
## Licence
Data is available for academic and commercial use, under a
[CC-BY-4.0 licence](http://creativecommons.org/licenses/by/4.0/legalcode).
EMBL-EBI expects attribution (e.g. in publications, services or products) for
any of its online services, databases or software in accordance with good
scientific practice.
If you make use of an AlphaFold prediction, please cite the following papers:
* [Jumper, J et al. Highly accurate protein structure prediction with
AlphaFold. Nature
(2021).](https://www.nature.com/articles/s41586-021-03819-2)
* [Varadi, M et al. AlphaFold Protein Structure Database: massively expanding
the structural coverage of protein-sequence space with high-accuracy models.
Nucleic Acids Research
(2021).](https://academic.oup.com/nar/advance-article/doi/10.1093/nar/gkab1061/6430488)
AlphaFold Data Copyright (2022) DeepMind Technologies Limited.
## Disclaimer
The AlphaFold Data and other information provided on this site is for
theoretical modelling only, caution should be exercised in its use. It is
provided 'as-is' without any warranty of any kind, whether expressed or implied.
For clarity, no warranty is given that use of the information shall not infringe
the rights of any third party. The information is not intended to be a
substitute for professional medical advice, diagnosis, or treatment, and does
not constitute medical or other professional advice.
## Format
Dataset file names start with a protein identifier of the form `AF-[a UniProt
accession]-F[a fragment number]`.
Three files are provided for each entry:
* **model_v4.cif** – contains the atomic coordinates for the predicted protein
structure, along with some metadata. Useful references for this file format
are the [ModelCIF](https://github.com/ihmwg/ModelCIF) and
[PDBx/mmCIF](https://mmcif.wwpdb.org) project sites.
* **confidence_v4.json** – contains a confidence metric output by AlphaFold
called pLDDT. This provides a number for each residue, indicating how
confident AlphaFold is in the *local* surrounding structure. pLDDT ranges
from 0 to 100, where 100 is most confident. This is also contained in the
CIF file.
* **predicted_aligned_error_v4.json** – contains a confidence metric output by
AlphaFold called PAE. This provides a number for every pair of residues,
which is lower when AlphaFold is more confident in the relative position of
the two residues. PAE is more suitable than pLDDT for judging confidence in
relative domain placements.
[See here](https://alphafold.ebi.ac.uk/faq#faq-7) for a description of the
format.
Predictions grouped by NCBI taxonomy ID are available as
`proteomes/proteome-tax_id-[TAX ID]-[SHARD ID]_v4.tar` within the same
bucket.
There are also two extra files stored in the bucket:
* `accession_ids.csv` – This file contains a list of all the UniProt
accessions that have predictions in AlphaFold DB. The file is in CSV format
and includes the following columns, separated by a comma:
* UniProt accession, e.g. A8H2R3
* First residue index (UniProt numbering), e.g. 1
* Last residue index (UniProt numbering), e.g. 199
* AlphaFold DB identifier, e.g. AF-A8H2R3-F1
* Latest version, e.g. 4
* `sequences.fasta` – This file contains sequences for all proteins in the
current database version in FASTA format. The identifier rows start with
">AFDB", followed by the AlphaFold DB identifier and the name of the
protein. The sequence rows contain the corresponding amino acid sequences.
Each sequence is on a single line, i.e. there is no wrapping.
## Creating a Google Cloud Account
Downloading from the Google Cloud Public Datasets (rather than from AFDB or 3D
Beacons) requires a Google Cloud account. See the
[Google Cloud get started](https://cloud.google.com/docs/get-started) page, and
explore the [free tier account usage limits](https://cloud.google.com/free).
**IMPORTANT: After the trial period has finished (90 days), to continue access,
you are required to upgrade to a billing account. While your free tier access
(including access to the Public Datasets storage bucket) continues, usage beyond
the free tier will incur costs – please familiarise yourself with the pricing
for the services that you use to avoid any surprises.**
1. Go to
[https://cloud.google.com/datasets](https://cloud.google.com/datasets).
2. Create an account:
1. Click "get started for free" in the top right corner.
2. Agree to all terms of service.
3. Follow the setup instructions. Note that a payment method is required,
but this will not be used unless you enable billing.
4. Access to the Google Cloud Public Datasets storage bucket is always at
no cost and you will have access to the
[free tier.](https://cloud.google.com/free/docs/gcp-free-tier#free-tier-usage-limits)
3. Set up a project:
1. In the top left corner, click the navigation menu (three horizontal bars
icon).
2. Select: "Cloud overview" -> "Dashboard".
3. In the top left corner there is a project menu bar (likely says "My
First Project"). Select this and a "Select a Project" box will appear.
4. To keep using this project, click "Cancel" at the bottom of the box.
5. To create a new project, click "New Project" at the top of the box:
1. Select a project name.
2. For location, if your organization has a Cloud account then select
this, otherwise leave as is.
4. Install `gsutil`:
1. Follow these
[instructions](https://cloud.google.com/storage/docs/gsutil_install).
## Accessing the dataset
The data is available from:
* GCS data bucket:
[gs://public-datasets-deepmind-alphafold-v4](https://console.cloud.google.com/storage/browser/public-datasets-deepmind-alphafold-v4)
## Bulk download
We don't recommend downloading the full dataset unless required for processing
with local computational resources, for example in an academic high performance
computing centre.
We estimate that a 1 Gbps internet connection will allow download of the full
database in roughly 2.5 days.
While we don’t know the exact nature of your computational infrastructure, below
are some suggested approaches for downloading the dataset. Please reach out to
[alphafold@deepmind.com](mailto:alphafold@deepmind.com) if you have any
questions.
The recommended way of downloading the whole database is by downloading
1,015,797 sharded proteome tar files using the command below. This is
significantly faster than downloading all of the individual files because of
large constant per-file latency.
```bash
gsutil -m cp -r gs://public-datasets-deepmind-alphafold-v4/proteomes/ .
```
You will then have to un-tar all of the proteomes and un-gzip all of the
individual files. Note that after un-taring, there will be about 644M files, so
make sure your filesystem can handle this.
### Storage Transfer Service
Some users might find the
[Storage Transfer Service](https://cloud.google.com/storage-transfer-service) a
convenient way to set up the transfer between this bucket and another bucket, or
another cloud service. *Using this service may incur costs*. Please check the
[pricing page](https://cloud.google.com/storage-transfer/pricing) for more
detail, particularly for transfers to other cloud services.
## Downloading subsets of the data
### AlphaFold Database search
For simple queries, for example by protein name, gene name or UniProt accession
you can use the main search bar on
[alphafold.ebi.ac.uk](https://alphafold.ebi.ac.uk).
### 3D Beacons
[3D-Beacons](https://3d-beacons.org) is an international collaboration of
protein structure data providers to create a federated network with unified data
access mechanisms. The 3D-Beacons platform allows users to retrieve coordinate
files and metadata of experimentally determined and theoretical protein models
from data providers such as AlphaFold DB.
More information about how to access AlphaFold predictions using 3D-Beacons is
available at
[3D-Beacons documentation](https://www.ebi.ac.uk/pdbe/pdbe-kb/3dbeacons/docs).
### Other premade species subsets
Downloads for some model organism proteomes, global health proteomes and
Swiss-Prot are available on the
[AFDB website](https://alphafold.ebi.ac.uk/download). These are generated from
[reference proteomes](https://www.uniprot.org/help/reference_proteome). If you
want other species, or *all* proteins for a particular species, please continue
reading.
We provide 1,015,797 sharded tar files for all species in
[gs://public-datasets-deepmind-alphafold-v4/proteomes/](https://console.cloud.google.com/storage/browser/public-datasets-deepmind-alphafold-v4/proteomes/).
We shard each proteome so that each shard contains at most 10,000 proteins
(which corresponds to 30,000 files per shard, since there are 3 files per
protein). To download a proteome of your choice, you have to do the following
steps:
1. Find the [NCBI taxonomy ID](https://www.ncbi.nlm.nih.gov/taxonomy)
(`[TAX_ID]`) of the species in question.
2. Run `gsutil -m cp
gs://public-datasets-deepmind-alphafold-v4/proteomes/proteome-tax_id-[TAX
ID]-*_v4.tar .` to download all shards for this proteome.
3. Un-tar all of the downloaded files and un-gzip all of the individual files.
### File manifests
Pre-made lists of files (manifests) are available at
[gs://public-datasets-deepmind-alphafold-v4/manifests](https://console.cloud.google.com/storage/browser/public-datasets-deepmind-alphafold-v4/manifests/).
Note that these filenames do not include the bucket prefix, but this can be
added once the files have been downloaded to your filesystem.
One can also define their own list of files, for example created by BigQuery
(see below). `gsutil` can be used to download these files with
```bash
cat [manifest file] | gsutil -m cp -I .
```
This will be much slower than downloading the tar files (grouped by species)
because each file has an associated overhead.
### BigQuery
**IMPORTANT: The
[free tier](https://cloud.google.com/bigquery/pricing#free-tier) of Google Cloud
comes with [BigQuery Sandbox](https://cloud.google.com/bigquery/docs/sandbox)
with 1 TB of free processed query data each month. Repeated queries within a
month could exceed this limit and if you have
[upgraded to a paid Cloud Billing account](https://cloud.google.com/free/docs/gcp-free-tier#how-to-upgrade)
you may be charged.**
**This should be sufficient for running a number of queries on the metadata
table, though the usage depends on the size of the columns queried and selected.
Please look at the
[BigQuery pricing page](https://cloud.google.com/bigquery/pricing) for more
information.**
**This is the user's responsibility so please ensure you keep track of your
billing settings and resource usage in the console.**
BigQuery provides a serverless and highly scalable analytics tool enabling SQL
queries over large datasets. The metadata for the UniProt dataset takes up ​​113
GiB and so can be challenging to process and analyse locally. The table name is:
* BigQuery metadata table:
[bigquery-public-data.deepmind_alphafold.metadata](https://console.cloud.google.com/bigquery?project=bigquery-public-data&ws=!1m5!1m4!4m3!1sbigquery-public-data!2sdeepmind_alphafold!3smetadata)
With BigQuery SQL you can do complex queries, e.g. find all high accuracy
predictions for a particular species, or even join on to other datasets, e.g. to
an experimental dataset by the `uniprotSequence`, or to the NCBI taxonomy by
`taxId`.
If you would find additional information in the metadata useful please file a
GitHub issue.
#### Setup
Follow the
[BigQuery Sandbox set up guide](https://cloud.google.com/bigquery/docs/sandbox).
#### Exploring the metadata
The column names and associated data types available can be found using the
following query.
```sql
SELECT column_name, data_type FROM bigquery-public-data.deepmind_alphafold.INFORMATION_SCHEMA.COLUMNS
WHERE table_name = 'metadata'
```
**Column name** | **Data type** | **Description**
---------------------- | --------------- | ---------------
allVersions | `ARRAY<INT64>` | An array of AFDB versions this prediction has had
entryId | `STRING` | The AFDB entry ID, e.g. "AF-Q1HGU3-F1"
fractionPlddtConfident | `FLOAT64` | Fraction of the residues in the prediction with pLDDT between 70 and 90
fractionPlddtLow | `FLOAT64` | Fraction of the residues in the prediction with pLDDT between 50 and 70
fractionPlddtVeryHigh | `FLOAT64` | Fraction of the residues in the prediction with pLDDT greater than 90
fractionPlddtVeryLow | `FLOAT64` | Fraction of the residues in the prediction with pLDDT less than 50
gene | `STRING` | The name of the gene if known, e.g. "COII"
geneSynonyms | `ARRAY<STRING>` | Additional synonyms for the gene
globalMetricValue | `FLOAT64` | The mean pLDDT of this prediction
isReferenceProteome | `BOOL` | Is this protein part of the reference proteome?
isReviewed | `BOOL` | Has this protein been reviewed, i.e. is it part of SwissProt?
latestVersion | `INT64` | The latest AFDB version for this prediction
modelCreatedDate | `DATE` | The date of creation for this entry, e.g. "2022-06-01"
organismCommonNames | `ARRAY<STRING>` | List of common organism names
organismScientificName | `STRING` | The scientific name of the organism
organismSynonyms | `ARRAY<STRING>` | List of synonyms for the organism
proteinFullNames | `ARRAY<STRING>` | Full names of the protein
proteinShortNames | `ARRAY<STRING>` | Short names of the protein
sequenceChecksum | `STRING` | [CRC64 hash](https://www.uniprot.org/help/checksum) of the sequence. Can be used for cheaper lookups.
sequenceVersionDate | `DATE` | Date when the sequence data was last modified in UniProt
taxId | `INT64` | NCBI taxonomy id of the originating species
uniprotAccession | `STRING` | Uniprot accession ID
uniprotDescription | `STRING` | The name recommended by the UniProt consortium
uniprotEnd | `INT64` | Number of the last residue in the entry relative to the UniProt entry. This is equal to the length of the protein unless we are dealing with protein fragments.
uniprotId | `STRING` | The Uniprot EntryName field
uniprotSequence | `STRING` | Amino acid sequence for this prediction
uniprotStart | `INT64` | Number of the first residue in the entry relative to the UniProt entry. This is 1 unless we are dealing with protein fragments.
#### Producing summary statistics
The following query gives the mean of the prediction confidence fractions per
species.
```sql
SELECT
organismScientificName AS name,
SUM(fractionPlddtVeryLow) / COUNT(fractionPlddtVeryLow) AS mean_plddt_very_low,
SUM(fractionPlddtLow) / COUNT(fractionPlddtLow) AS mean_plddt_low,
SUM(fractionPlddtConfident) / COUNT(fractionPlddtConfident) AS mean_plddt_confident,
SUM(fractionPlddtVeryHigh) / COUNT(fractionPlddtVeryHigh) AS mean_plddt_very_high,
COUNT(organismScientificName) AS num_predictions
FROM bigquery-public-data.deepmind_alphafold.metadata
GROUP by name
ORDER BY num_predictions DESC;
```
#### Producing lists of files
We expect that the most important use for the metadata will be to create subsets
of proteins according to various criteria, so that users can choose to only copy
a subset of the 214M proteins that exist in the dataset. An example query is
given below:
```sql
with file_rows AS (
with file_cols AS (
SELECT
CONCAT(entryID, '-model_v4.cif') as m,
CONCAT(entryID, '-predicted_aligned_error_v4.json') as p
FROM bigquery-public-data.deepmind_alphafold.metadata
WHERE organismScientificName = "Homo sapiens"
AND (fractionPlddtVeryHigh + fractionPlddtConfident) > 0.5
)
SELECT * FROM file_cols UNPIVOT (files for filetype in (m, p))
)
SELECT CONCAT('gs://public-datasets-deepmind-alphafold-v4/', files) as files
from file_rows
```
In this case, the list has been filtered to only include proteins from *Homo
sapiens* for which over half the residues are confident or better (>70 pLDDT).
This creates a table with one column "files", where each row is the cloud
location of one of the two file types that has been provided for each protein.
There is an additional `confidence_v4.json` file which contains the
per-residue pLDDT. This information is already in the CIF file but may be
preferred if only this information is required.
This allows users to bulk download the exact proteins they need, without having
to download the entire dataset. Other columns may also be used to select subsets
of proteins, and we point the user to the
[BigQuery documentation](https://cloud.google.com/bigquery/docs) to understand
other ways to filter for their desired protein lists. Likewise, the
documentation should be followed to download these file subsets locally, as the
most appropriate approach will depend on the filesize. Note that it may be
easier to download large files using [Colab](https://colab.research.google.com/)
(e.g. pandas to_csv).
#### Previous versions
Previous versions of AFDB will remain available at
[gs://public-datasets-deepmind-alphafold](https://console.cloud.google.com/storage/browser/public-datasets-deepmind-alphafold)
to enable reproducible research. We recommend using the latest version (v4).
\ No newline at end of file
alphafold/__init__.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""An implementation of the inference pipeline of AlphaFold v2.0."""
alphafold/common/__init__.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Common data types and constants used within Alphafold."""
alphafold/common/confidence.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Functions for processing confidence metrics."""
import json
from typing import Dict, Optional, Tuple
import numpy as np
import scipy.special
def compute_plddt(logits: np.ndarray) -> np.ndarray:
"""Computes per-residue pLDDT from logits.
Args:
logits: [num_res, num_bins] output from the PredictedLDDTHead.
Returns:
plddt: [num_res] per-residue pLDDT.
"""
num_bins = logits.shape[-1]
bin_width = 1.0 / num_bins
bin_centers = np.arange(start=0.5 * bin_width, stop=1.0, step=bin_width)
probs = scipy.special.softmax(logits, axis=-1)
predicted_lddt_ca = np.sum(probs * bin_centers[None, :], axis=-1)
return predicted_lddt_ca * 100
def _confidence_category(score: float) -> str:
"""Categorizes pLDDT into: disordered (D), low (L), medium (M), high (H)."""
if 0 <= score < 50:
return 'D'
if 50 <= score < 70:
return 'L'
elif 70 <= score < 90:
return 'M'
elif 90 <= score <= 100:
return 'H'
else:
raise ValueError(f'Invalid pLDDT score {score}')
def confidence_json(plddt: np.ndarray) -> str:
"""Returns JSON with confidence score and category for every residue.
Args:
plddt: Per-residue confidence metric data.
Returns:
String with a formatted JSON.
Raises:
ValueError: If `plddt` has a rank different than 1.
"""
if plddt.ndim != 1:
raise ValueError(f'The plddt array must be rank 1, got: {plddt.shape}.')
confidence = {
'residueNumber': list(range(1, len(plddt) + 1)),
'confidenceScore': [round(float(s), 2) for s in plddt],
'confidenceCategory': [_confidence_category(s) for s in plddt],
}
return json.dumps(confidence, indent=None, separators=(',', ':'))
def _calculate_bin_centers(breaks: np.ndarray):
"""Gets the bin centers from the bin edges.
Args:
breaks: [num_bins - 1] the error bin edges.
Returns:
bin_centers: [num_bins] the error bin centers.
"""
step = (breaks[1] - breaks[0])
# Add half-step to get the center
bin_centers = breaks + step / 2
# Add a catch-all bin at the end.
bin_centers = np.concatenate([bin_centers, [bin_centers[-1] + step]],
axis=0)
return bin_centers
def _calculate_expected_aligned_error(
alignment_confidence_breaks: np.ndarray,
aligned_distance_error_probs: np.ndarray) -> Tuple[np.ndarray, np.ndarray]:
"""Calculates expected aligned distance errors for every pair of residues.
Args:
alignment_confidence_breaks: [num_bins - 1] the error bin edges.
aligned_distance_error_probs: [num_res, num_res, num_bins] the predicted
probs for each error bin, for each pair of residues.
Returns:
predicted_aligned_error: [num_res, num_res] the expected aligned distance
error for each pair of residues.
max_predicted_aligned_error: The maximum predicted error possible.
"""
bin_centers = _calculate_bin_centers(alignment_confidence_breaks)
# Tuple of expected aligned distance error and max possible error.
return (np.sum(aligned_distance_error_probs * bin_centers, axis=-1),
np.asarray(bin_centers[-1]))
def compute_predicted_aligned_error(
logits: np.ndarray,
breaks: np.ndarray) -> Dict[str, np.ndarray]:
"""Computes aligned confidence metrics from logits.
Args:
logits: [num_res, num_res, num_bins] the logits output from
PredictedAlignedErrorHead.
breaks: [num_bins - 1] the error bin edges.
Returns:
aligned_confidence_probs: [num_res, num_res, num_bins] the predicted
aligned error probabilities over bins for each residue pair.
predicted_aligned_error: [num_res, num_res] the expected aligned distance
error for each pair of residues.
max_predicted_aligned_error: The maximum predicted error possible.
"""
aligned_confidence_probs = scipy.special.softmax(
logits,
axis=-1)
predicted_aligned_error, max_predicted_aligned_error = (
_calculate_expected_aligned_error(
alignment_confidence_breaks=breaks,
aligned_distance_error_probs=aligned_confidence_probs))
return {
'aligned_confidence_probs': aligned_confidence_probs,
'predicted_aligned_error': predicted_aligned_error,
'max_predicted_aligned_error': max_predicted_aligned_error,
}
def pae_json(pae: np.ndarray, max_pae: float) -> str:
"""Returns the PAE in the same format as is used in the AFDB.
Note that the values are presented as floats to 1 decimal place, whereas AFDB
returns integer values.
Args:
pae: The n_res x n_res PAE array.
max_pae: The maximum possible PAE value.
Returns:
PAE output format as a JSON string.
"""
# Check the PAE array is the correct shape.
if pae.ndim != 2 or pae.shape[0] != pae.shape[1]:
raise ValueError(f'PAE must be a square matrix, got {pae.shape}')
# Round the predicted aligned errors to 1 decimal place.
rounded_errors = np.round(pae.astype(np.float64), decimals=1)
formatted_output = [{
'predicted_aligned_error': rounded_errors.tolist(),
'max_predicted_aligned_error': max_pae,
}]
return json.dumps(formatted_output, indent=None, separators=(',', ':'))
def predicted_tm_score(
logits: np.ndarray,
breaks: np.ndarray,
residue_weights: Optional[np.ndarray] = None,
asym_id: Optional[np.ndarray] = None,
interface: bool = False) -> np.ndarray:
"""Computes predicted TM alignment or predicted interface TM alignment score.
Args:
logits: [num_res, num_res, num_bins] the logits output from
PredictedAlignedErrorHead.
breaks: [num_bins] the error bins.
residue_weights: [num_res] the per residue weights to use for the
expectation.
asym_id: [num_res] the asymmetric unit ID - the chain ID. Only needed for
ipTM calculation, i.e. when interface=True.
interface: If True, interface predicted TM score is computed.
Returns:
ptm_score: The predicted TM alignment or the predicted iTM score.
"""
# residue_weights has to be in [0, 1], but can be floating-point, i.e. the
# exp. resolved head's probability.
if residue_weights is None:
residue_weights = np.ones(logits.shape[0])
bin_centers = _calculate_bin_centers(breaks)
num_res = int(np.sum(residue_weights))
# Clip num_res to avoid negative/undefined d0.
clipped_num_res = max(num_res, 19)
# Compute d_0(num_res) as defined by TM-score, eqn. (5) in Yang & Skolnick
# "Scoring function for automated assessment of protein structure template
# quality", 2004: http://zhanglab.ccmb.med.umich.edu/papers/2004_3.pdf
d0 = 1.24 * (clipped_num_res - 15) ** (1./3) - 1.8
# Convert logits to probs.
probs = scipy.special.softmax(logits, axis=-1)
# TM-Score term for every bin.
tm_per_bin = 1. / (1 + np.square(bin_centers) / np.square(d0))
# E_distances tm(distance).
predicted_tm_term = np.sum(probs * tm_per_bin, axis=-1)
pair_mask = np.ones(shape=(num_res, num_res), dtype=bool)
if interface:
pair_mask *= asym_id[:, None] != asym_id[None, :]
predicted_tm_term *= pair_mask
pair_residue_weights = pair_mask * (
residue_weights[None, :] * residue_weights[:, None])
normed_residue_mask = pair_residue_weights / (1e-8 + np.sum(
pair_residue_weights, axis=-1, keepdims=True))
per_alignment = np.sum(predicted_tm_term * normed_residue_mask, axis=-1)
return np.asarray(per_alignment[(per_alignment * residue_weights).argmax()])
alphafold/common/confidence_test.py 0 → 100644
View file @ eb93322b
# Copyright 2023 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Test confidence metrics."""
from absl.testing import absltest
from alphafold.common import confidence
import numpy as np
class ConfidenceTest(absltest.TestCase):
def test_pae_json(self):
pae = np.array([[0.01, 13.12345], [20.0987, 0.0]])
pae_json = confidence.pae_json(pae=pae, max_pae=31.75)
self.assertEqual(
pae_json, '[{"predicted_aligned_error":[[0.0,13.1],[20.1,0.0]],'
'"max_predicted_aligned_error":31.75}]')
def test_confidence_json(self):
plddt = np.array([42, 42.42])
confidence_json = confidence.confidence_json(plddt=plddt)
print(confidence_json)
self.assertEqual(
confidence_json,
('{"residueNumber":[1,2],'
'"confidenceScore":[42.0,42.42],'
'"confidenceCategory":["D","D"]}'),
)
if __name__ == '__main__':
absltest.main()
alphafold/common/mmcif_metadata.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""mmCIF metadata."""
from typing import Mapping, Sequence
from alphafold import version
import numpy as np
_DISCLAIMER = """ALPHAFOLD DATA, COPYRIGHT (2021) DEEPMIND TECHNOLOGIES LIMITED.
THE INFORMATION PROVIDED IS THEORETICAL MODELLING ONLY AND CAUTION SHOULD BE
EXERCISED IN ITS USE. IT IS PROVIDED "AS-IS" WITHOUT ANY WARRANTY OF ANY KIND,
WHETHER EXPRESSED OR IMPLIED. NO WARRANTY IS GIVEN THAT USE OF THE INFORMATION
SHALL NOT INFRINGE THE RIGHTS OF ANY THIRD PARTY. DISCLAIMER: THE INFORMATION IS
NOT INTENDED TO BE A SUBSTITUTE FOR PROFESSIONAL MEDICAL ADVICE, DIAGNOSIS, OR
TREATMENT, AND DOES NOT CONSTITUTE MEDICAL OR OTHER PROFESSIONAL ADVICE. IT IS
AVAILABLE FOR ACADEMIC AND COMMERCIAL PURPOSES, UNDER CC-BY 4.0 LICENCE."""
# Authors of the Nature methods paper we reference in the mmCIF.
_MMCIF_PAPER_AUTHORS = (
'Jumper, John',
'Evans, Richard',
'Pritzel, Alexander',
'Green, Tim',
'Figurnov, Michael',
'Ronneberger, Olaf',
'Tunyasuvunakool, Kathryn',
'Bates, Russ',
'Zidek, Augustin',
'Potapenko, Anna',
'Bridgland, Alex',
'Meyer, Clemens',
'Kohl, Simon A. A.',
'Ballard, Andrew J.',
'Cowie, Andrew',
'Romera-Paredes, Bernardino',
'Nikolov, Stanislav',
'Jain, Rishub',
'Adler, Jonas',
'Back, Trevor',
'Petersen, Stig',
'Reiman, David',
'Clancy, Ellen',
'Zielinski, Michal',
'Steinegger, Martin',
'Pacholska, Michalina',
'Berghammer, Tamas',
'Silver, David',
'Vinyals, Oriol',
'Senior, Andrew W.',
'Kavukcuoglu, Koray',
'Kohli, Pushmeet',
'Hassabis, Demis',
)
# Authors of the mmCIF - we set them to be equal to the authors of the paper.
_MMCIF_AUTHORS = _MMCIF_PAPER_AUTHORS
def add_metadata_to_mmcif(
old_cif: Mapping[str, Sequence[str]], model_type: str
) -> Mapping[str, Sequence[str]]:
"""Adds AlphaFold metadata in the given mmCIF."""
cif = {}
# ModelCIF conformation dictionary.
cif['_audit_conform.dict_name'] = ['mmcif_ma.dic']
cif['_audit_conform.dict_version'] = ['1.3.9']
cif['_audit_conform.dict_location'] = [
'https://raw.githubusercontent.com/ihmwg/ModelCIF/master/dist/'
'mmcif_ma.dic'
]
# License and disclaimer.
cif['_pdbx_data_usage.id'] = ['1', '2']
cif['_pdbx_data_usage.type'] = ['license', 'disclaimer']
cif['_pdbx_data_usage.details'] = [
'Data in this file is available under a CC-BY-4.0 license.',
_DISCLAIMER,
]
cif['_pdbx_data_usage.url'] = [
'https://creativecommons.org/licenses/by/4.0/',
'?',
]
cif['_pdbx_data_usage.name'] = ['CC-BY-4.0', '?']
# Structure author details.
cif['_audit_author.name'] = []
cif['_audit_author.pdbx_ordinal'] = []
for author_index, author_name in enumerate(_MMCIF_AUTHORS, start=1):
cif['_audit_author.name'].append(author_name)
cif['_audit_author.pdbx_ordinal'].append(str(author_index))
# Paper author details.
cif['_citation_author.citation_id'] = []
cif['_citation_author.name'] = []
cif['_citation_author.ordinal'] = []
for author_index, author_name in enumerate(_MMCIF_PAPER_AUTHORS, start=1):
cif['_citation_author.citation_id'].append('primary')
cif['_citation_author.name'].append(author_name)
cif['_citation_author.ordinal'].append(str(author_index))
# Paper citation details.
cif['_citation.id'] = ['primary']
cif['_citation.title'] = [
'Highly accurate protein structure prediction with AlphaFold'
]
cif['_citation.journal_full'] = ['Nature']
cif['_citation.journal_volume'] = ['596']
cif['_citation.page_first'] = ['583']
cif['_citation.page_last'] = ['589']
cif['_citation.year'] = ['2021']
cif['_citation.journal_id_ASTM'] = ['NATUAS']
cif['_citation.country'] = ['UK']
cif['_citation.journal_id_ISSN'] = ['0028-0836']
cif['_citation.journal_id_CSD'] = ['0006']
cif['_citation.book_publisher'] = ['?']
cif['_citation.pdbx_database_id_PubMed'] = ['34265844']
cif['_citation.pdbx_database_id_DOI'] = ['10.1038/s41586-021-03819-2']
# Type of data in the dataset including data used in the model generation.
cif['_ma_data.id'] = ['1']
cif['_ma_data.name'] = ['Model']
cif['_ma_data.content_type'] = ['model coordinates']
# Description of number of instances for each entity.
cif['_ma_target_entity_instance.asym_id'] = old_cif['_struct_asym.id']
cif['_ma_target_entity_instance.entity_id'] = old_cif[
'_struct_asym.entity_id'
]
cif['_ma_target_entity_instance.details'] = ['.'] * len(
cif['_ma_target_entity_instance.entity_id']
)
# Details about the target entities.
cif['_ma_target_entity.entity_id'] = cif[
'_ma_target_entity_instance.entity_id'
]
cif['_ma_target_entity.data_id'] = ['1'] * len(
cif['_ma_target_entity.entity_id']
)
cif['_ma_target_entity.origin'] = ['.'] * len(
cif['_ma_target_entity.entity_id']
)
# Details of the models being deposited.
cif['_ma_model_list.ordinal_id'] = ['1']
cif['_ma_model_list.model_id'] = ['1']
cif['_ma_model_list.model_group_id'] = ['1']
cif['_ma_model_list.model_name'] = ['Top ranked model']
cif['_ma_model_list.model_group_name'] = [
f'AlphaFold {model_type} v{version.__version__} model'
]
cif['_ma_model_list.data_id'] = ['1']
cif['_ma_model_list.model_type'] = ['Ab initio model']
# Software used.
cif['_software.pdbx_ordinal'] = ['1']
cif['_software.name'] = ['AlphaFold']
cif['_software.version'] = [f'v{version.__version__}']
cif['_software.type'] = ['package']
cif['_software.description'] = ['Structure prediction']
cif['_software.classification'] = ['other']
cif['_software.date'] = ['?']
# Collection of software into groups.
cif['_ma_software_group.ordinal_id'] = ['1']
cif['_ma_software_group.group_id'] = ['1']
cif['_ma_software_group.software_id'] = ['1']
# Method description to conform with ModelCIF.
cif['_ma_protocol_step.ordinal_id'] = ['1', '2', '3']
cif['_ma_protocol_step.protocol_id'] = ['1', '1', '1']
cif['_ma_protocol_step.step_id'] = ['1', '2', '3']
cif['_ma_protocol_step.method_type'] = [
'coevolution MSA',
'template search',
'modeling',
]
# Details of the metrics use to assess model confidence.
cif['_ma_qa_metric.id'] = ['1', '2']
cif['_ma_qa_metric.name'] = ['pLDDT', 'pLDDT']
# Accepted values are distance, energy, normalised score, other, zscore.
cif['_ma_qa_metric.type'] = ['pLDDT', 'pLDDT']
cif['_ma_qa_metric.mode'] = ['global', 'local']
cif['_ma_qa_metric.software_group_id'] = ['1', '1']
# Global model confidence metric value.
cif['_ma_qa_metric_global.ordinal_id'] = ['1']
cif['_ma_qa_metric_global.model_id'] = ['1']
cif['_ma_qa_metric_global.metric_id'] = ['1']
global_plddt = np.mean(
[float(v) for v in old_cif['_atom_site.B_iso_or_equiv']]
)
cif['_ma_qa_metric_global.metric_value'] = [f'{global_plddt:.2f}']
cif['_atom_type.symbol'] = sorted(set(old_cif['_atom_site.type_symbol']))
return cif
alphafold/common/protein.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Protein data type."""
import collections
import dataclasses
import functools
import io
from typing import Any, Dict, List, Mapping, Optional, Tuple
from alphafold.common import mmcif_metadata
from alphafold.common import residue_constants
from Bio.PDB import MMCIFParser
from Bio.PDB import PDBParser
from Bio.PDB.mmcifio import MMCIFIO
from Bio.PDB.Structure import Structure
import numpy as np
FeatureDict = Mapping[str, np.ndarray]
ModelOutput = Mapping[str, Any] # Is a nested dict.
# Complete sequence of chain IDs supported by the PDB format.
PDB_CHAIN_IDS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789'
PDB_MAX_CHAINS = len(PDB_CHAIN_IDS) # := 62.
# Data to fill the _chem_comp table when writing mmCIFs.
_CHEM_COMP: Mapping[str, Tuple[Tuple[str, str], ...]] = {
'L-peptide linking': (
('ALA', 'ALANINE'),
('ARG', 'ARGININE'),
('ASN', 'ASPARAGINE'),
('ASP', 'ASPARTIC ACID'),
('CYS', 'CYSTEINE'),
('GLN', 'GLUTAMINE'),
('GLU', 'GLUTAMIC ACID'),
('HIS', 'HISTIDINE'),
('ILE', 'ISOLEUCINE'),
('LEU', 'LEUCINE'),
('LYS', 'LYSINE'),
('MET', 'METHIONINE'),
('PHE', 'PHENYLALANINE'),
('PRO', 'PROLINE'),
('SER', 'SERINE'),
('THR', 'THREONINE'),
('TRP', 'TRYPTOPHAN'),
('TYR', 'TYROSINE'),
('VAL', 'VALINE'),
),
'peptide linking': (('GLY', 'GLYCINE'),),
}
@dataclasses.dataclass(frozen=True)
class Protein:
"""Protein structure representation."""
# Cartesian coordinates of atoms in angstroms. The atom types correspond to
# residue_constants.atom_types, i.e. the first three are N, CA, CB.
atom_positions: np.ndarray # [num_res, num_atom_type, 3]
# Amino-acid type for each residue represented as an integer between 0 and
# 20, where 20 is 'X'.
aatype: np.ndarray # [num_res]
# Binary float mask to indicate presence of a particular atom. 1.0 if an atom
# is present and 0.0 if not. This should be used for loss masking.
atom_mask: np.ndarray # [num_res, num_atom_type]
# Residue index as used in PDB. It is not necessarily continuous or 0-indexed.
residue_index: np.ndarray # [num_res]
# 0-indexed number corresponding to the chain in the protein that this residue
# belongs to.
chain_index: np.ndarray # [num_res]
# B-factors, or temperature factors, of each residue (in sq. angstroms units),
# representing the displacement of the residue from its ground truth mean
# value.
b_factors: np.ndarray # [num_res, num_atom_type]
def __post_init__(self):
if len(np.unique(self.chain_index)) > PDB_MAX_CHAINS:
raise ValueError(
f'Cannot build an instance with more than {PDB_MAX_CHAINS} chains '
'because these cannot be written to PDB format.')
def _from_bio_structure(
structure: Structure, chain_id: Optional[str] = None
) -> Protein:
"""Takes a Biopython structure and creates a `Protein` instance.
WARNING: All non-standard residue types will be converted into UNK. All
non-standard atoms will be ignored.
Args:
structure: Structure from the Biopython library.
chain_id: If chain_id is specified (e.g. A), then only that chain is parsed.
Otherwise all chains are parsed.
Returns:
A new `Protein` created from the structure contents.
Raises:
ValueError: If the number of models included in the structure is not 1.
ValueError: If insertion code is detected at a residue.
"""
models = list(structure.get_models())
if len(models) != 1:
raise ValueError(
'Only single model PDBs/mmCIFs are supported. Found'
f' {len(models)} models.'
)
model = models[0]
atom_positions = []
aatype = []
atom_mask = []
residue_index = []
chain_ids = []
b_factors = []
for chain in model:
if chain_id is not None and chain.id != chain_id:
continue
for res in chain:
if res.id[2] != ' ':
raise ValueError(
f'PDB/mmCIF contains an insertion code at chain {chain.id} and'
f' residue index {res.id[1]}. These are not supported.'
)
res_shortname = residue_constants.restype_3to1.get(res.resname, 'X')
restype_idx = residue_constants.restype_order.get(
res_shortname, residue_constants.restype_num)
pos = np.zeros((residue_constants.atom_type_num, 3))
mask = np.zeros((residue_constants.atom_type_num,))
res_b_factors = np.zeros((residue_constants.atom_type_num,))
for atom in res:
if atom.name not in residue_constants.atom_types:
continue
pos[residue_constants.atom_order[atom.name]] = atom.coord
mask[residue_constants.atom_order[atom.name]] = 1.
res_b_factors[residue_constants.atom_order[atom.name]] = atom.bfactor
if np.sum(mask) < 0.5:
# If no known atom positions are reported for the residue then skip it.
continue
aatype.append(restype_idx)
atom_positions.append(pos)
atom_mask.append(mask)
residue_index.append(res.id[1])
chain_ids.append(chain.id)
b_factors.append(res_b_factors)
# Chain IDs are usually characters so map these to ints.
unique_chain_ids = np.unique(chain_ids)
chain_id_mapping = {cid: n for n, cid in enumerate(unique_chain_ids)}
chain_index = np.array([chain_id_mapping[cid] for cid in chain_ids])
return Protein(
atom_positions=np.array(atom_positions),
atom_mask=np.array(atom_mask),
aatype=np.array(aatype),
residue_index=np.array(residue_index),
chain_index=chain_index,
b_factors=np.array(b_factors))
def from_pdb_string(pdb_str: str, chain_id: Optional[str] = None) -> Protein:
"""Takes a PDB string and constructs a `Protein` object.
WARNING: All non-standard residue types will be converted into UNK. All
non-standard atoms will be ignored.
Args:
pdb_str: The contents of the pdb file
chain_id: If chain_id is specified (e.g. A), then only that chain is parsed.
Otherwise all chains are parsed.
Returns:
A new `Protein` parsed from the pdb contents.
"""
with io.StringIO(pdb_str) as pdb_fh:
parser = PDBParser(QUIET=True)
structure = parser.get_structure(id='none', file=pdb_fh)
return _from_bio_structure(structure, chain_id)
def from_mmcif_string(
mmcif_str: str, chain_id: Optional[str] = None
) -> Protein:
"""Takes a mmCIF string and constructs a `Protein` object.
WARNING: All non-standard residue types will be converted into UNK. All
non-standard atoms will be ignored.
Args:
mmcif_str: The contents of the mmCIF file
chain_id: If chain_id is specified (e.g. A), then only that chain is parsed.
Otherwise all chains are parsed.
Returns:
A new `Protein` parsed from the mmCIF contents.
"""
with io.StringIO(mmcif_str) as mmcif_fh:
parser = MMCIFParser(QUIET=True)
structure = parser.get_structure(structure_id='none', filename=mmcif_fh)
return _from_bio_structure(structure, chain_id)
def _chain_end(atom_index, end_resname, chain_name, residue_index) -> str:
chain_end = 'TER'
return (f'{chain_end:<6}{atom_index:>5} {end_resname:>3} '
f'{chain_name:>1}{residue_index:>4}')
def to_pdb(prot: Protein) -> str:
"""Converts a `Protein` instance to a PDB string.
Args:
prot: The protein to convert to PDB.
Returns:
PDB string.
"""
restypes = residue_constants.restypes + ['X']
res_1to3 = lambda r: residue_constants.restype_1to3.get(restypes[r], 'UNK')
atom_types = residue_constants.atom_types
pdb_lines = []
atom_mask = prot.atom_mask
aatype = prot.aatype
atom_positions = prot.atom_positions
residue_index = prot.residue_index.astype(np.int32)
chain_index = prot.chain_index.astype(np.int32)
b_factors = prot.b_factors
if np.any(aatype > residue_constants.restype_num):
raise ValueError('Invalid aatypes.')
# Construct a mapping from chain integer indices to chain ID strings.
chain_ids = {}
for i in np.unique(chain_index): # np.unique gives sorted output.
if i >= PDB_MAX_CHAINS:
raise ValueError(
f'The PDB format supports at most {PDB_MAX_CHAINS} chains.')
chain_ids[i] = PDB_CHAIN_IDS[i]
pdb_lines.append('MODEL 1')
atom_index = 1
last_chain_index = chain_index[0]
# Add all atom sites.
for i in range(aatype.shape[0]):
# Close the previous chain if in a multichain PDB.
if last_chain_index != chain_index[i]:
pdb_lines.append(_chain_end(
atom_index, res_1to3(aatype[i - 1]), chain_ids[chain_index[i - 1]],
residue_index[i - 1]))
last_chain_index = chain_index[i]
atom_index += 1 # Atom index increases at the TER symbol.
res_name_3 = res_1to3(aatype[i])
for atom_name, pos, mask, b_factor in zip(
atom_types, atom_positions[i], atom_mask[i], b_factors[i]):
if mask < 0.5:
continue
record_type = 'ATOM'
name = atom_name if len(atom_name) == 4 else f' {atom_name}'
alt_loc = ''
insertion_code = ''
occupancy = 1.00
element = atom_name[0] # Protein supports only C, N, O, S, this works.
charge = ''
# PDB is a columnar format, every space matters here!
atom_line = (f'{record_type:<6}{atom_index:>5} {name:<4}{alt_loc:>1}'
f'{res_name_3:>3} {chain_ids[chain_index[i]]:>1}'
f'{residue_index[i]:>4}{insertion_code:>1} '
f'{pos[0]:>8.3f}{pos[1]:>8.3f}{pos[2]:>8.3f}'
f'{occupancy:>6.2f}{b_factor:>6.2f} '
f'{element:>2}{charge:>2}')
pdb_lines.append(atom_line)
atom_index += 1
# Close the final chain.
pdb_lines.append(_chain_end(atom_index, res_1to3(aatype[-1]),
chain_ids[chain_index[-1]], residue_index[-1]))
pdb_lines.append('ENDMDL')
pdb_lines.append('END')
# Pad all lines to 80 characters.
pdb_lines = [line.ljust(80) for line in pdb_lines]
return '\n'.join(pdb_lines) + '\n' # Add terminating newline.
def ideal_atom_mask(prot: Protein) -> np.ndarray:
"""Computes an ideal atom mask.
`Protein.atom_mask` typically is defined according to the atoms that are
reported in the PDB. This function computes a mask according to heavy atoms
that should be present in the given sequence of amino acids.
Args:
prot: `Protein` whose fields are `numpy.ndarray` objects.
Returns:
An ideal atom mask.
"""
return residue_constants.STANDARD_ATOM_MASK[prot.aatype]
def from_prediction(
features: FeatureDict,
result: ModelOutput,
b_factors: Optional[np.ndarray] = None,
remove_leading_feature_dimension: bool = True) -> Protein:
"""Assembles a protein from a prediction.
Args:
features: Dictionary holding model inputs.
result: Dictionary holding model outputs.
b_factors: (Optional) B-factors to use for the protein.
remove_leading_feature_dimension: Whether to remove the leading dimension
of the `features` values.
Returns:
A protein instance.
"""
fold_output = result['structure_module']
def _maybe_remove_leading_dim(arr: np.ndarray) -> np.ndarray:
return arr[0] if remove_leading_feature_dimension else arr
if 'asym_id' in features:
chain_index = _maybe_remove_leading_dim(features['asym_id'])
else:
chain_index = np.zeros_like(_maybe_remove_leading_dim(features['aatype']))
if b_factors is None:
b_factors = np.zeros_like(fold_output['final_atom_mask'])
return Protein(
aatype=_maybe_remove_leading_dim(features['aatype']),
atom_positions=fold_output['final_atom_positions'],
atom_mask=fold_output['final_atom_mask'],
residue_index=_maybe_remove_leading_dim(features['residue_index']) + 1,
chain_index=chain_index,
b_factors=b_factors)
def to_mmcif(
prot: Protein,
file_id: str,
model_type: str,
) -> str:
"""Converts a `Protein` instance to an mmCIF string.
WARNING 1: The _entity_poly_seq is filled with unknown (UNK) residues for any
missing residue indices in the range from min(1, min(residue_index)) to
max(residue_index). E.g. for a protein object with positions for residues
2 (MET), 3 (LYS), 6 (GLY), this method would set the _entity_poly_seq to:
1 UNK
2 MET
3 LYS
4 UNK
5 UNK
6 GLY
This is done to preserve the residue numbering.
WARNING 2: Converting ground truth mmCIF file to Protein and then back to
mmCIF using this method will convert all non-standard residue types to UNK.
If you need this behaviour, you need to store more mmCIF metadata in the
Protein object (e.g. all fields except for the _atom_site loop).
WARNING 3: Converting ground truth mmCIF file to Protein and then back to
mmCIF using this method will not retain the original chain indices.
WARNING 4: In case of multiple identical chains, they are assigned different
`_atom_site.label_entity_id` values.
Args:
prot: A protein to convert to mmCIF string.
file_id: The file ID (usually the PDB ID) to be used in the mmCIF.
model_type: 'Multimer' or 'Monomer'.
Returns:
A valid mmCIF string.
Raises:
ValueError: If aminoacid types array contains entries with too many protein
types.
"""
atom_mask = prot.atom_mask
aatype = prot.aatype
atom_positions = prot.atom_positions
residue_index = prot.residue_index.astype(np.int32)
chain_index = prot.chain_index.astype(np.int32)
b_factors = prot.b_factors
# Construct a mapping from chain integer indices to chain ID strings.
chain_ids = {}
# We count unknown residues as protein residues.
for entity_id in np.unique(chain_index): # np.unique gives sorted output.
chain_ids[entity_id] = _int_id_to_str_id(entity_id + 1)
mmcif_dict = collections.defaultdict(list)
mmcif_dict['data_'] = file_id.upper()
mmcif_dict['_entry.id'] = file_id.upper()
label_asym_id_to_entity_id = {}
# Entity and chain information.
for entity_id, chain_id in chain_ids.items():
# Add all chain information to the _struct_asym table.
label_asym_id_to_entity_id[str(chain_id)] = str(entity_id)
mmcif_dict['_struct_asym.id'].append(chain_id)
mmcif_dict['_struct_asym.entity_id'].append(str(entity_id))
# Add information about the entity to the _entity_poly table.
mmcif_dict['_entity_poly.entity_id'].append(str(entity_id))
mmcif_dict['_entity_poly.type'].append(residue_constants.PROTEIN_CHAIN)
mmcif_dict['_entity_poly.pdbx_strand_id'].append(chain_id)
# Generate the _entity table.
mmcif_dict['_entity.id'].append(str(entity_id))
mmcif_dict['_entity.type'].append(residue_constants.POLYMER_CHAIN)
# Add the residues to the _entity_poly_seq table.
for entity_id, (res_ids, aas) in _get_entity_poly_seq(
aatype, residue_index, chain_index
).items():
for res_id, aa in zip(res_ids, aas):
mmcif_dict['_entity_poly_seq.entity_id'].append(str(entity_id))
mmcif_dict['_entity_poly_seq.num'].append(str(res_id))
mmcif_dict['_entity_poly_seq.mon_id'].append(
residue_constants.resnames[aa]
)
# Populate the chem comp table.
for chem_type, chem_comp in _CHEM_COMP.items():
for chem_id, chem_name in chem_comp:
mmcif_dict['_chem_comp.id'].append(chem_id)
mmcif_dict['_chem_comp.type'].append(chem_type)
mmcif_dict['_chem_comp.name'].append(chem_name)
# Add all atom sites.
atom_index = 1
for i in range(aatype.shape[0]):
res_name_3 = residue_constants.resnames[aatype[i]]
if aatype[i] <= len(residue_constants.restypes):
atom_names = residue_constants.atom_types
else:
raise ValueError(
'Amino acid types array contains entries with too many protein types.'
)
for atom_name, pos, mask, b_factor in zip(
atom_names, atom_positions[i], atom_mask[i], b_factors[i]
):
if mask < 0.5:
continue
type_symbol = residue_constants.atom_id_to_type(atom_name)
mmcif_dict['_atom_site.group_PDB'].append('ATOM')
mmcif_dict['_atom_site.id'].append(str(atom_index))
mmcif_dict['_atom_site.type_symbol'].append(type_symbol)
mmcif_dict['_atom_site.label_atom_id'].append(atom_name)
mmcif_dict['_atom_site.label_alt_id'].append('.')
mmcif_dict['_atom_site.label_comp_id'].append(res_name_3)
mmcif_dict['_atom_site.label_asym_id'].append(chain_ids[chain_index[i]])
mmcif_dict['_atom_site.label_entity_id'].append(
label_asym_id_to_entity_id[chain_ids[chain_index[i]]]
)
mmcif_dict['_atom_site.label_seq_id'].append(str(residue_index[i]))
mmcif_dict['_atom_site.pdbx_PDB_ins_code'].append('.')
mmcif_dict['_atom_site.Cartn_x'].append(f'{pos[0]:.3f}')
mmcif_dict['_atom_site.Cartn_y'].append(f'{pos[1]:.3f}')
mmcif_dict['_atom_site.Cartn_z'].append(f'{pos[2]:.3f}')
mmcif_dict['_atom_site.occupancy'].append('1.00')
mmcif_dict['_atom_site.B_iso_or_equiv'].append(f'{b_factor:.2f}')
mmcif_dict['_atom_site.auth_seq_id'].append(str(residue_index[i]))
mmcif_dict['_atom_site.auth_asym_id'].append(chain_ids[chain_index[i]])
mmcif_dict['_atom_site.pdbx_PDB_model_num'].append('1')
atom_index += 1
metadata_dict = mmcif_metadata.add_metadata_to_mmcif(mmcif_dict, model_type)
mmcif_dict.update(metadata_dict)
return _create_mmcif_string(mmcif_dict)
@functools.lru_cache(maxsize=256)
def _int_id_to_str_id(num: int) -> str:
"""Encodes a number as a string, using reverse spreadsheet style naming.
Args:
num: A positive integer.
Returns:
A string that encodes the positive integer using reverse spreadsheet style,
naming e.g. 1 = A, 2 = B, ..., 27 = AA, 28 = BA, 29 = CA, ... This is the
usual way to encode chain IDs in mmCIF files.
"""
if num <= 0:
raise ValueError(f'Only positive integers allowed, got {num}.')
num = num - 1 # 1-based indexing.
output = []
while num >= 0:
output.append(chr(num % 26 + ord('A')))
num = num // 26 - 1
return ''.join(output)
def _get_entity_poly_seq(
aatypes: np.ndarray, residue_indices: np.ndarray, chain_indices: np.ndarray
) -> Dict[int, Tuple[List[int], List[int]]]:
"""Constructs gapless residue index and aatype lists for each chain.
Args:
aatypes: A numpy array with aatypes.
residue_indices: A numpy array with residue indices.
chain_indices: A numpy array with chain indices.
Returns:
A dictionary mapping chain indices to a tuple with list of residue indices
and a list of aatypes. Missing residues are filled with UNK residue type.
"""
if (
aatypes.shape[0] != residue_indices.shape[0]
or aatypes.shape[0] != chain_indices.shape[0]
):
raise ValueError(
'aatypes, residue_indices, chain_indices must have the same length.'
)
# Group the present residues by chain index.
present = collections.defaultdict(list)
for chain_id, res_id, aa in zip(chain_indices, residue_indices, aatypes):
present[chain_id].append((res_id, aa))
# Add any missing residues (from 1 to the first residue and for any gaps).
entity_poly_seq = {}
for chain_id, present_residues in present.items():
present_residue_indices = set([x[0] for x in present_residues])
min_res_id = min(present_residue_indices) # Could be negative.
max_res_id = max(present_residue_indices)
new_residue_indices = []
new_aatypes = []
present_index = 0
for i in range(min(1, min_res_id), max_res_id + 1):
new_residue_indices.append(i)
if i in present_residue_indices:
new_aatypes.append(present_residues[present_index][1])
present_index += 1
else:
new_aatypes.append(20) # Unknown amino acid type.
entity_poly_seq[chain_id] = (new_residue_indices, new_aatypes)
return entity_poly_seq
def _create_mmcif_string(mmcif_dict: Dict[str, Any]) -> str:
"""Converts mmCIF dictionary into mmCIF string."""
mmcifio = MMCIFIO()
mmcifio.set_dict(mmcif_dict)
with io.StringIO() as file_handle:
mmcifio.save(file_handle)
return file_handle.getvalue()
alphafold/common/protein_test.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Tests for protein."""
import os
from absl.testing import absltest
from absl.testing import parameterized
from alphafold.common import protein
from alphafold.common import residue_constants
import numpy as np
# Internal import (7716).
TEST_DATA_DIR = 'alphafold/common/testdata/'
class ProteinTest(parameterized.TestCase):
def _check_shapes(self, prot, num_res):
"""Check that the processed shapes are correct."""
num_atoms = residue_constants.atom_type_num
self.assertEqual((num_res, num_atoms, 3), prot.atom_positions.shape)
self.assertEqual((num_res,), prot.aatype.shape)
self.assertEqual((num_res, num_atoms), prot.atom_mask.shape)
self.assertEqual((num_res,), prot.residue_index.shape)
self.assertEqual((num_res,), prot.chain_index.shape)
self.assertEqual((num_res, num_atoms), prot.b_factors.shape)
@parameterized.named_parameters(
dict(testcase_name='chain_A',
pdb_file='2rbg.pdb', chain_id='A', num_res=282, num_chains=1),
dict(testcase_name='chain_B',
pdb_file='2rbg.pdb', chain_id='B', num_res=282, num_chains=1),
dict(testcase_name='multichain',
pdb_file='2rbg.pdb', chain_id=None, num_res=564, num_chains=2))
def test_from_pdb_str(self, pdb_file, chain_id, num_res, num_chains):
pdb_file = os.path.join(absltest.get_default_test_srcdir(), TEST_DATA_DIR,
pdb_file)
with open(pdb_file) as f:
pdb_string = f.read()
prot = protein.from_pdb_string(pdb_string, chain_id)
self._check_shapes(prot, num_res)
self.assertGreaterEqual(prot.aatype.min(), 0)
# Allow equal since unknown restypes have index equal to restype_num.
self.assertLessEqual(prot.aatype.max(), residue_constants.restype_num)
self.assertLen(np.unique(prot.chain_index), num_chains)
def test_to_pdb(self):
with open(
os.path.join(absltest.get_default_test_srcdir(), TEST_DATA_DIR,
'2rbg.pdb')) as f:
pdb_string = f.read()
prot = protein.from_pdb_string(pdb_string)
pdb_string_reconstr = protein.to_pdb(prot)
for line in pdb_string_reconstr.splitlines():
self.assertLen(line, 80)
prot_reconstr = protein.from_pdb_string(pdb_string_reconstr)
np.testing.assert_array_equal(prot_reconstr.aatype, prot.aatype)
np.testing.assert_array_almost_equal(
prot_reconstr.atom_positions, prot.atom_positions)
np.testing.assert_array_almost_equal(
prot_reconstr.atom_mask, prot.atom_mask)
np.testing.assert_array_equal(
prot_reconstr.residue_index, prot.residue_index)
np.testing.assert_array_equal(
prot_reconstr.chain_index, prot.chain_index)
np.testing.assert_array_almost_equal(
prot_reconstr.b_factors, prot.b_factors)
@parameterized.named_parameters(
dict(
testcase_name='glucagon',
pdb_file='glucagon.pdb',
model_type='Monomer',
),
dict(testcase_name='7bui', pdb_file='5nmu.pdb', model_type='Multimer'),
)
def test_to_mmcif(self, pdb_file, model_type):
with open(
os.path.join(
absltest.get_default_test_srcdir(), TEST_DATA_DIR, pdb_file
)
) as f:
pdb_string = f.read()
prot = protein.from_pdb_string(pdb_string)
file_id = 'test'
mmcif_string = protein.to_mmcif(prot, file_id, model_type)
prot_reconstr = protein.from_mmcif_string(mmcif_string)
np.testing.assert_array_equal(prot_reconstr.aatype, prot.aatype)
np.testing.assert_array_almost_equal(
prot_reconstr.atom_positions, prot.atom_positions
)
np.testing.assert_array_almost_equal(
prot_reconstr.atom_mask, prot.atom_mask
)
np.testing.assert_array_equal(
prot_reconstr.residue_index, prot.residue_index
)
np.testing.assert_array_equal(prot_reconstr.chain_index, prot.chain_index)
np.testing.assert_array_almost_equal(
prot_reconstr.b_factors, prot.b_factors
)
def test_ideal_atom_mask(self):
with open(
os.path.join(
absltest.get_default_test_srcdir(), TEST_DATA_DIR, '2rbg.pdb'
)
) as f:
pdb_string = f.read()
prot = protein.from_pdb_string(pdb_string)
ideal_mask = protein.ideal_atom_mask(prot)
non_ideal_residues = set([102] + list(range(127, 286)))
for i, (res, atom_mask) in enumerate(
zip(prot.residue_index, prot.atom_mask)
):
if res in non_ideal_residues:
self.assertFalse(np.all(atom_mask == ideal_mask[i]), msg=f'{res}')
else:
self.assertTrue(np.all(atom_mask == ideal_mask[i]), msg=f'{res}')
def test_too_many_chains(self):
num_res = protein.PDB_MAX_CHAINS + 1
num_atom_type = residue_constants.atom_type_num
with self.assertRaises(ValueError):
_ = protein.Protein(
atom_positions=np.random.random([num_res, num_atom_type, 3]),
aatype=np.random.randint(0, 21, [num_res]),
atom_mask=np.random.randint(0, 2, [num_res]).astype(np.float32),
residue_index=np.arange(1, num_res+1),
chain_index=np.arange(num_res),
b_factors=np.random.uniform(1, 100, [num_res]))
if __name__ == '__main__':
absltest.main()
alphafold/common/residue_constants.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Constants used in AlphaFold."""
import collections
import functools
import os
from typing import Final, List, Mapping, Tuple
import numpy as np
import tree
# Internal import (35fd).
# Distance from one CA to next CA [trans configuration: omega = 180].
ca_ca = 3.80209737096
# Format: The list for each AA type contains chi1, chi2, chi3, chi4 in
# this order (or a relevant subset from chi1 onwards). ALA and GLY don't have
# chi angles so their chi angle lists are empty.
chi_angles_atoms = {
'ALA': [],
# Chi5 in arginine is always 0 +- 5 degrees, so ignore it.
'ARG': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD'],
['CB', 'CG', 'CD', 'NE'], ['CG', 'CD', 'NE', 'CZ']],
'ASN': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'OD1']],
'ASP': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'OD1']],
'CYS': [['N', 'CA', 'CB', 'SG']],
'GLN': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD'],
['CB', 'CG', 'CD', 'OE1']],
'GLU': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD'],
['CB', 'CG', 'CD', 'OE1']],
'GLY': [],
'HIS': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'ND1']],
'ILE': [['N', 'CA', 'CB', 'CG1'], ['CA', 'CB', 'CG1', 'CD1']],
'LEU': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD1']],
'LYS': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD'],
['CB', 'CG', 'CD', 'CE'], ['CG', 'CD', 'CE', 'NZ']],
'MET': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'SD'],
['CB', 'CG', 'SD', 'CE']],
'PHE': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD1']],
'PRO': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD']],
'SER': [['N', 'CA', 'CB', 'OG']],
'THR': [['N', 'CA', 'CB', 'OG1']],
'TRP': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD1']],
'TYR': [['N', 'CA', 'CB', 'CG'], ['CA', 'CB', 'CG', 'CD1']],
'VAL': [['N', 'CA', 'CB', 'CG1']],
}
# If chi angles given in fixed-length array, this matrix determines how to mask
# them for each AA type. The order is as per restype_order (see below).
chi_angles_mask = [
[0.0, 0.0, 0.0, 0.0], # ALA
[1.0, 1.0, 1.0, 1.0], # ARG
[1.0, 1.0, 0.0, 0.0], # ASN
[1.0, 1.0, 0.0, 0.0], # ASP
[1.0, 0.0, 0.0, 0.0], # CYS
[1.0, 1.0, 1.0, 0.0], # GLN
[1.0, 1.0, 1.0, 0.0], # GLU
[0.0, 0.0, 0.0, 0.0], # GLY
[1.0, 1.0, 0.0, 0.0], # HIS
[1.0, 1.0, 0.0, 0.0], # ILE
[1.0, 1.0, 0.0, 0.0], # LEU
[1.0, 1.0, 1.0, 1.0], # LYS
[1.0, 1.0, 1.0, 0.0], # MET
[1.0, 1.0, 0.0, 0.0], # PHE
[1.0, 1.0, 0.0, 0.0], # PRO
[1.0, 0.0, 0.0, 0.0], # SER
[1.0, 0.0, 0.0, 0.0], # THR
[1.0, 1.0, 0.0, 0.0], # TRP
[1.0, 1.0, 0.0, 0.0], # TYR
[1.0, 0.0, 0.0, 0.0], # VAL
]
# The following chi angles are pi periodic: they can be rotated by a multiple
# of pi without affecting the structure.
chi_pi_periodic = [
[0.0, 0.0, 0.0, 0.0], # ALA
[0.0, 0.0, 0.0, 0.0], # ARG
[0.0, 0.0, 0.0, 0.0], # ASN
[0.0, 1.0, 0.0, 0.0], # ASP
[0.0, 0.0, 0.0, 0.0], # CYS
[0.0, 0.0, 0.0, 0.0], # GLN
[0.0, 0.0, 1.0, 0.0], # GLU
[0.0, 0.0, 0.0, 0.0], # GLY
[0.0, 0.0, 0.0, 0.0], # HIS
[0.0, 0.0, 0.0, 0.0], # ILE
[0.0, 0.0, 0.0, 0.0], # LEU
[0.0, 0.0, 0.0, 0.0], # LYS
[0.0, 0.0, 0.0, 0.0], # MET
[0.0, 1.0, 0.0, 0.0], # PHE
[0.0, 0.0, 0.0, 0.0], # PRO
[0.0, 0.0, 0.0, 0.0], # SER
[0.0, 0.0, 0.0, 0.0], # THR
[0.0, 0.0, 0.0, 0.0], # TRP
[0.0, 1.0, 0.0, 0.0], # TYR
[0.0, 0.0, 0.0, 0.0], # VAL
[0.0, 0.0, 0.0, 0.0], # UNK
]
# Atoms positions relative to the 8 rigid groups, defined by the pre-omega, phi,
# psi and chi angles:
# 0: 'backbone group',
# 1: 'pre-omega-group', (empty)
# 2: 'phi-group', (currently empty, because it defines only hydrogens)
# 3: 'psi-group',
# 4,5,6,7: 'chi1,2,3,4-group'
# The atom positions are relative to the axis-end-atom of the corresponding
# rotation axis. The x-axis is in direction of the rotation axis, and the y-axis
# is defined such that the dihedral-angle-defining atom (the last entry in
# chi_angles_atoms above) is in the xy-plane (with a positive y-coordinate).
# format: [atomname, group_idx, rel_position]
rigid_group_atom_positions = {
'ALA': [
['N', 0, (-0.525, 1.363, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, -0.000, -0.000)],
['CB', 0, (-0.529, -0.774, -1.205)],
['O', 3, (0.627, 1.062, 0.000)],
],
'ARG': [
['N', 0, (-0.524, 1.362, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, -0.000, -0.000)],
['CB', 0, (-0.524, -0.778, -1.209)],
['O', 3, (0.626, 1.062, 0.000)],
['CG', 4, (0.616, 1.390, -0.000)],
['CD', 5, (0.564, 1.414, 0.000)],
['NE', 6, (0.539, 1.357, -0.000)],
['NH1', 7, (0.206, 2.301, 0.000)],
['NH2', 7, (2.078, 0.978, -0.000)],
['CZ', 7, (0.758, 1.093, -0.000)],
],
'ASN': [
['N', 0, (-0.536, 1.357, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, -0.000, -0.000)],
['CB', 0, (-0.531, -0.787, -1.200)],
['O', 3, (0.625, 1.062, 0.000)],
['CG', 4, (0.584, 1.399, 0.000)],
['ND2', 5, (0.593, -1.188, 0.001)],
['OD1', 5, (0.633, 1.059, 0.000)],
],
'ASP': [
['N', 0, (-0.525, 1.362, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.527, 0.000, -0.000)],
['CB', 0, (-0.526, -0.778, -1.208)],
['O', 3, (0.626, 1.062, -0.000)],
['CG', 4, (0.593, 1.398, -0.000)],
['OD1', 5, (0.610, 1.091, 0.000)],
['OD2', 5, (0.592, -1.101, -0.003)],
],
'CYS': [
['N', 0, (-0.522, 1.362, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.524, 0.000, 0.000)],
['CB', 0, (-0.519, -0.773, -1.212)],
['O', 3, (0.625, 1.062, -0.000)],
['SG', 4, (0.728, 1.653, 0.000)],
],
'GLN': [
['N', 0, (-0.526, 1.361, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, 0.000, 0.000)],
['CB', 0, (-0.525, -0.779, -1.207)],
['O', 3, (0.626, 1.062, -0.000)],
['CG', 4, (0.615, 1.393, 0.000)],
['CD', 5, (0.587, 1.399, -0.000)],
['NE2', 6, (0.593, -1.189, -0.001)],
['OE1', 6, (0.634, 1.060, 0.000)],
],
'GLU': [
['N', 0, (-0.528, 1.361, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, -0.000, -0.000)],
['CB', 0, (-0.526, -0.781, -1.207)],
['O', 3, (0.626, 1.062, 0.000)],
['CG', 4, (0.615, 1.392, 0.000)],
['CD', 5, (0.600, 1.397, 0.000)],
['OE1', 6, (0.607, 1.095, -0.000)],
['OE2', 6, (0.589, -1.104, -0.001)],
],
'GLY': [
['N', 0, (-0.572, 1.337, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.517, -0.000, -0.000)],
['O', 3, (0.626, 1.062, -0.000)],
],
'HIS': [
['N', 0, (-0.527, 1.360, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, 0.000, 0.000)],
['CB', 0, (-0.525, -0.778, -1.208)],
['O', 3, (0.625, 1.063, 0.000)],
['CG', 4, (0.600, 1.370, -0.000)],
['CD2', 5, (0.889, -1.021, 0.003)],
['ND1', 5, (0.744, 1.160, -0.000)],
['CE1', 5, (2.030, 0.851, 0.002)],
['NE2', 5, (2.145, -0.466, 0.004)],
],
'ILE': [
['N', 0, (-0.493, 1.373, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.527, -0.000, -0.000)],
['CB', 0, (-0.536, -0.793, -1.213)],
['O', 3, (0.627, 1.062, -0.000)],
['CG1', 4, (0.534, 1.437, -0.000)],
['CG2', 4, (0.540, -0.785, -1.199)],
['CD1', 5, (0.619, 1.391, 0.000)],
],
'LEU': [
['N', 0, (-0.520, 1.363, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, -0.000, -0.000)],
['CB', 0, (-0.522, -0.773, -1.214)],
['O', 3, (0.625, 1.063, -0.000)],
['CG', 4, (0.678, 1.371, 0.000)],
['CD1', 5, (0.530, 1.430, -0.000)],
['CD2', 5, (0.535, -0.774, 1.200)],
],
'LYS': [
['N', 0, (-0.526, 1.362, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, 0.000, 0.000)],
['CB', 0, (-0.524, -0.778, -1.208)],
['O', 3, (0.626, 1.062, -0.000)],
['CG', 4, (0.619, 1.390, 0.000)],
['CD', 5, (0.559, 1.417, 0.000)],
['CE', 6, (0.560, 1.416, 0.000)],
['NZ', 7, (0.554, 1.387, 0.000)],
],
'MET': [
['N', 0, (-0.521, 1.364, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, 0.000, 0.000)],
['CB', 0, (-0.523, -0.776, -1.210)],
['O', 3, (0.625, 1.062, -0.000)],
['CG', 4, (0.613, 1.391, -0.000)],
['SD', 5, (0.703, 1.695, 0.000)],
['CE', 6, (0.320, 1.786, -0.000)],
],
'PHE': [
['N', 0, (-0.518, 1.363, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.524, 0.000, -0.000)],
['CB', 0, (-0.525, -0.776, -1.212)],
['O', 3, (0.626, 1.062, -0.000)],
['CG', 4, (0.607, 1.377, 0.000)],
['CD1', 5, (0.709, 1.195, -0.000)],
['CD2', 5, (0.706, -1.196, 0.000)],
['CE1', 5, (2.102, 1.198, -0.000)],
['CE2', 5, (2.098, -1.201, -0.000)],
['CZ', 5, (2.794, -0.003, -0.001)],
],
'PRO': [
['N', 0, (-0.566, 1.351, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.527, -0.000, 0.000)],
['CB', 0, (-0.546, -0.611, -1.293)],
['O', 3, (0.621, 1.066, 0.000)],
['CG', 4, (0.382, 1.445, 0.0)],
# ['CD', 5, (0.427, 1.440, 0.0)],
['CD', 5, (0.477, 1.424, 0.0)], # manually made angle 2 degrees larger
],
'SER': [
['N', 0, (-0.529, 1.360, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, -0.000, -0.000)],
['CB', 0, (-0.518, -0.777, -1.211)],
['O', 3, (0.626, 1.062, -0.000)],
['OG', 4, (0.503, 1.325, 0.000)],
],
'THR': [
['N', 0, (-0.517, 1.364, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.526, 0.000, -0.000)],
['CB', 0, (-0.516, -0.793, -1.215)],
['O', 3, (0.626, 1.062, 0.000)],
['CG2', 4, (0.550, -0.718, -1.228)],
['OG1', 4, (0.472, 1.353, 0.000)],
],
'TRP': [
['N', 0, (-0.521, 1.363, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.525, -0.000, 0.000)],
['CB', 0, (-0.523, -0.776, -1.212)],
['O', 3, (0.627, 1.062, 0.000)],
['CG', 4, (0.609, 1.370, -0.000)],
['CD1', 5, (0.824, 1.091, 0.000)],
['CD2', 5, (0.854, -1.148, -0.005)],
['CE2', 5, (2.186, -0.678, -0.007)],
['CE3', 5, (0.622, -2.530, -0.007)],
['NE1', 5, (2.140, 0.690, -0.004)],
['CH2', 5, (3.028, -2.890, -0.013)],
['CZ2', 5, (3.283, -1.543, -0.011)],
['CZ3', 5, (1.715, -3.389, -0.011)],
],
'TYR': [
['N', 0, (-0.522, 1.362, 0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.524, -0.000, -0.000)],
['CB', 0, (-0.522, -0.776, -1.213)],
['O', 3, (0.627, 1.062, -0.000)],
['CG', 4, (0.607, 1.382, -0.000)],
['CD1', 5, (0.716, 1.195, -0.000)],
['CD2', 5, (0.713, -1.194, -0.001)],
['CE1', 5, (2.107, 1.200, -0.002)],
['CE2', 5, (2.104, -1.201, -0.003)],
['OH', 5, (4.168, -0.002, -0.005)],
['CZ', 5, (2.791, -0.001, -0.003)],
],
'VAL': [
['N', 0, (-0.494, 1.373, -0.000)],
['CA', 0, (0.000, 0.000, 0.000)],
['C', 0, (1.527, -0.000, -0.000)],
['CB', 0, (-0.533, -0.795, -1.213)],
['O', 3, (0.627, 1.062, -0.000)],
['CG1', 4, (0.540, 1.429, -0.000)],
['CG2', 4, (0.533, -0.776, 1.203)],
],
}
# A list of atoms (excluding hydrogen) for each AA type. PDB naming convention.
residue_atoms = {
'ALA': ['C', 'CA', 'CB', 'N', 'O'],
'ARG': ['C', 'CA', 'CB', 'CG', 'CD', 'CZ', 'N', 'NE', 'O', 'NH1', 'NH2'],
'ASP': ['C', 'CA', 'CB', 'CG', 'N', 'O', 'OD1', 'OD2'],
'ASN': ['C', 'CA', 'CB', 'CG', 'N', 'ND2', 'O', 'OD1'],
'CYS': ['C', 'CA', 'CB', 'N', 'O', 'SG'],
'GLU': ['C', 'CA', 'CB', 'CG', 'CD', 'N', 'O', 'OE1', 'OE2'],
'GLN': ['C', 'CA', 'CB', 'CG', 'CD', 'N', 'NE2', 'O', 'OE1'],
'GLY': ['C', 'CA', 'N', 'O'],
'HIS': ['C', 'CA', 'CB', 'CG', 'CD2', 'CE1', 'N', 'ND1', 'NE2', 'O'],
'ILE': ['C', 'CA', 'CB', 'CG1', 'CG2', 'CD1', 'N', 'O'],
'LEU': ['C', 'CA', 'CB', 'CG', 'CD1', 'CD2', 'N', 'O'],
'LYS': ['C', 'CA', 'CB', 'CG', 'CD', 'CE', 'N', 'NZ', 'O'],
'MET': ['C', 'CA', 'CB', 'CG', 'CE', 'N', 'O', 'SD'],
'PHE': ['C', 'CA', 'CB', 'CG', 'CD1', 'CD2', 'CE1', 'CE2', 'CZ', 'N', 'O'],
'PRO': ['C', 'CA', 'CB', 'CG', 'CD', 'N', 'O'],
'SER': ['C', 'CA', 'CB', 'N', 'O', 'OG'],
'THR': ['C', 'CA', 'CB', 'CG2', 'N', 'O', 'OG1'],
'TRP': ['C', 'CA', 'CB', 'CG', 'CD1', 'CD2', 'CE2', 'CE3', 'CZ2', 'CZ3',
'CH2', 'N', 'NE1', 'O'],
'TYR': ['C', 'CA', 'CB', 'CG', 'CD1', 'CD2', 'CE1', 'CE2', 'CZ', 'N', 'O',
'OH'],
'VAL': ['C', 'CA', 'CB', 'CG1', 'CG2', 'N', 'O']
}
# Naming swaps for ambiguous atom names.
# Due to symmetries in the amino acids the naming of atoms is ambiguous in
# 4 of the 20 amino acids.
# (The LDDT paper lists 7 amino acids as ambiguous, but the naming ambiguities
# in LEU, VAL and ARG can be resolved by using the 3d constellations of
# the 'ambiguous' atoms and their neighbours)
residue_atom_renaming_swaps = {
'ASP': {'OD1': 'OD2'},
'GLU': {'OE1': 'OE2'},
'PHE': {'CD1': 'CD2', 'CE1': 'CE2'},
'TYR': {'CD1': 'CD2', 'CE1': 'CE2'},
}
# Van der Waals radii [Angstroem] of the atoms (from Wikipedia)
van_der_waals_radius = {
'C': 1.7,
'N': 1.55,
'O': 1.52,
'S': 1.8,
}
Bond = collections.namedtuple(
'Bond', ['atom1_name', 'atom2_name', 'length', 'stddev'])
BondAngle = collections.namedtuple(
'BondAngle',
['atom1_name', 'atom2_name', 'atom3name', 'angle_rad', 'stddev'])
@functools.lru_cache(maxsize=None)
def load_stereo_chemical_props() -> Tuple[Mapping[str, List[Bond]],
Mapping[str, List[Bond]],
Mapping[str, List[BondAngle]]]:
"""Load stereo_chemical_props.txt into a nice structure.
Load literature values for bond lengths and bond angles and translate
bond angles into the length of the opposite edge of the triangle
("residue_virtual_bonds").
Returns:
residue_bonds: Dict that maps resname -> list of Bond tuples.
residue_virtual_bonds: Dict that maps resname -> list of Bond tuples.
residue_bond_angles: Dict that maps resname -> list of BondAngle tuples.
"""
stereo_chemical_props_path = os.path.join(
os.path.dirname(os.path.abspath(__file__)), 'stereo_chemical_props.txt'
)
with open(stereo_chemical_props_path, 'rt') as f:
stereo_chemical_props = f.read()
lines_iter = iter(stereo_chemical_props.splitlines())
# Load bond lengths.
residue_bonds = {}
next(lines_iter) # Skip header line.
for line in lines_iter:
if line.strip() == '-':
break
bond, resname, length, stddev = line.split()
atom1, atom2 = bond.split('-')
if resname not in residue_bonds:
residue_bonds[resname] = []
residue_bonds[resname].append(
Bond(atom1, atom2, float(length), float(stddev)))
residue_bonds['UNK'] = []
# Load bond angles.
residue_bond_angles = {}
next(lines_iter) # Skip empty line.
next(lines_iter) # Skip header line.
for line in lines_iter:
if line.strip() == '-':
break
bond, resname, angle_degree, stddev_degree = line.split()
atom1, atom2, atom3 = bond.split('-')
if resname not in residue_bond_angles:
residue_bond_angles[resname] = []
residue_bond_angles[resname].append(
BondAngle(atom1, atom2, atom3,
float(angle_degree) / 180. * np.pi,
float(stddev_degree) / 180. * np.pi))
residue_bond_angles['UNK'] = []
def make_bond_key(atom1_name, atom2_name):
"""Unique key to lookup bonds."""
return '-'.join(sorted([atom1_name, atom2_name]))
# Translate bond angles into distances ("virtual bonds").
residue_virtual_bonds = {}
for resname, bond_angles in residue_bond_angles.items():
# Create a fast lookup dict for bond lengths.
bond_cache = {}
for b in residue_bonds[resname]:
bond_cache[make_bond_key(b.atom1_name, b.atom2_name)] = b
residue_virtual_bonds[resname] = []
for ba in bond_angles:
bond1 = bond_cache[make_bond_key(ba.atom1_name, ba.atom2_name)]
bond2 = bond_cache[make_bond_key(ba.atom2_name, ba.atom3name)]
# Compute distance between atom1 and atom3 using the law of cosines
# c^2 = a^2 + b^2 - 2ab*cos(gamma).
gamma = ba.angle_rad
length = np.sqrt(bond1.length**2 + bond2.length**2
- 2 * bond1.length * bond2.length * np.cos(gamma))
# Propagation of uncertainty assuming uncorrelated errors.
dl_outer = 0.5 / length
dl_dgamma = (2 * bond1.length * bond2.length * np.sin(gamma)) * dl_outer
dl_db1 = (2 * bond1.length - 2 * bond2.length * np.cos(gamma)) * dl_outer
dl_db2 = (2 * bond2.length - 2 * bond1.length * np.cos(gamma)) * dl_outer
stddev = np.sqrt((dl_dgamma * ba.stddev)**2 +
(dl_db1 * bond1.stddev)**2 +
(dl_db2 * bond2.stddev)**2)
residue_virtual_bonds[resname].append(
Bond(ba.atom1_name, ba.atom3name, length, stddev))
return (residue_bonds,
residue_virtual_bonds,
residue_bond_angles)
# Between-residue bond lengths for general bonds (first element) and for Proline
# (second element).
between_res_bond_length_c_n = [1.329, 1.341]
between_res_bond_length_stddev_c_n = [0.014, 0.016]
# Between-residue cos_angles.
between_res_cos_angles_c_n_ca = [-0.5203, 0.0353] # degrees: 121.352 +- 2.315
between_res_cos_angles_ca_c_n = [-0.4473, 0.0311] # degrees: 116.568 +- 1.995
# This mapping is used when we need to store atom data in a format that requires
# fixed atom data size for every residue (e.g. a numpy array).
atom_types = [
'N', 'CA', 'C', 'CB', 'O', 'CG', 'CG1', 'CG2', 'OG', 'OG1', 'SG', 'CD',
'CD1', 'CD2', 'ND1', 'ND2', 'OD1', 'OD2', 'SD', 'CE', 'CE1', 'CE2', 'CE3',
'NE', 'NE1', 'NE2', 'OE1', 'OE2', 'CH2', 'NH1', 'NH2', 'OH', 'CZ', 'CZ2',
'CZ3', 'NZ', 'OXT'
]
atom_order = {atom_type: i for i, atom_type in enumerate(atom_types)}
atom_type_num = len(atom_types) # := 37.
# A compact atom encoding with 14 columns
# pylint: disable=line-too-long
# pylint: disable=bad-whitespace
restype_name_to_atom14_names = {
'ALA': ['N', 'CA', 'C', 'O', 'CB', '', '', '', '', '', '', '', '', ''],
'ARG': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD', 'NE', 'CZ', 'NH1', 'NH2', '', '', ''],
'ASN': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'OD1', 'ND2', '', '', '', '', '', ''],
'ASP': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'OD1', 'OD2', '', '', '', '', '', ''],
'CYS': ['N', 'CA', 'C', 'O', 'CB', 'SG', '', '', '', '', '', '', '', ''],
'GLN': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD', 'OE1', 'NE2', '', '', '', '', ''],
'GLU': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD', 'OE1', 'OE2', '', '', '', '', ''],
'GLY': ['N', 'CA', 'C', 'O', '', '', '', '', '', '', '', '', '', ''],
'HIS': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'ND1', 'CD2', 'CE1', 'NE2', '', '', '', ''],
'ILE': ['N', 'CA', 'C', 'O', 'CB', 'CG1', 'CG2', 'CD1', '', '', '', '', '', ''],
'LEU': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD1', 'CD2', '', '', '', '', '', ''],
'LYS': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD', 'CE', 'NZ', '', '', '', '', ''],
'MET': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'SD', 'CE', '', '', '', '', '', ''],
'PHE': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD1', 'CD2', 'CE1', 'CE2', 'CZ', '', '', ''],
'PRO': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD', '', '', '', '', '', '', ''],
'SER': ['N', 'CA', 'C', 'O', 'CB', 'OG', '', '', '', '', '', '', '', ''],
'THR': ['N', 'CA', 'C', 'O', 'CB', 'OG1', 'CG2', '', '', '', '', '', '', ''],
'TRP': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD1', 'CD2', 'NE1', 'CE2', 'CE3', 'CZ2', 'CZ3', 'CH2'],
'TYR': ['N', 'CA', 'C', 'O', 'CB', 'CG', 'CD1', 'CD2', 'CE1', 'CE2', 'CZ', 'OH', '', ''],
'VAL': ['N', 'CA', 'C', 'O', 'CB', 'CG1', 'CG2', '', '', '', '', '', '', ''],
'UNK': ['', '', '', '', '', '', '', '', '', '', '', '', '', ''],
}
# pylint: enable=line-too-long
# pylint: enable=bad-whitespace
# This is the standard residue order when coding AA type as a number.
# Reproduce it by taking 3-letter AA codes and sorting them alphabetically.
restypes = [
'A', 'R', 'N', 'D', 'C', 'Q', 'E', 'G', 'H', 'I', 'L', 'K', 'M', 'F', 'P',
'S', 'T', 'W', 'Y', 'V'
]
restype_order = {restype: i for i, restype in enumerate(restypes)}
restype_num = len(restypes) # := 20.
unk_restype_index = restype_num # Catch-all index for unknown restypes.
restypes_with_x = restypes + ['X']
restype_order_with_x = {restype: i for i, restype in enumerate(restypes_with_x)}
def sequence_to_onehot(
sequence: str,
mapping: Mapping[str, int],
map_unknown_to_x: bool = False) -> np.ndarray:
"""Maps the given sequence into a one-hot encoded matrix.
Args:
sequence: An amino acid sequence.
mapping: A dictionary mapping amino acids to integers.
map_unknown_to_x: If True, any amino acid that is not in the mapping will be
mapped to the unknown amino acid 'X'. If the mapping doesn't contain
amino acid 'X', an error will be thrown. If False, any amino acid not in
the mapping will throw an error.
Returns:
A numpy array of shape (seq_len, num_unique_aas) with one-hot encoding of
the sequence.
Raises:
ValueError: If the mapping doesn't contain values from 0 to
num_unique_aas - 1 without any gaps.
"""
num_entries = max(mapping.values()) + 1
if sorted(set(mapping.values())) != list(range(num_entries)):
raise ValueError('The mapping must have values from 0 to num_unique_aas-1 '
'without any gaps. Got: %s' % sorted(mapping.values()))
one_hot_arr = np.zeros((len(sequence), num_entries), dtype=np.int32)
for aa_index, aa_type in enumerate(sequence):
if map_unknown_to_x:
if aa_type.isalpha() and aa_type.isupper():
aa_id = mapping.get(aa_type, mapping['X'])
else:
raise ValueError(f'Invalid character in the sequence: {aa_type}')
else:
aa_id = mapping[aa_type]
one_hot_arr[aa_index, aa_id] = 1
return one_hot_arr
restype_1to3 = {
'A': 'ALA',
'R': 'ARG',
'N': 'ASN',
'D': 'ASP',
'C': 'CYS',
'Q': 'GLN',
'E': 'GLU',
'G': 'GLY',
'H': 'HIS',
'I': 'ILE',
'L': 'LEU',
'K': 'LYS',
'M': 'MET',
'F': 'PHE',
'P': 'PRO',
'S': 'SER',
'T': 'THR',
'W': 'TRP',
'Y': 'TYR',
'V': 'VAL',
}
PROTEIN_CHAIN: Final[str] = 'polypeptide(L)'
POLYMER_CHAIN: Final[str] = 'polymer'
def atom_id_to_type(atom_id: str) -> str:
"""Convert atom ID to atom type, works only for standard protein residues.
Args:
atom_id: Atom ID to be converted.
Returns:
String corresponding to atom type.
Raises:
ValueError: If atom ID not recognized.
"""
if atom_id.startswith('C'):
return 'C'
elif atom_id.startswith('N'):
return 'N'
elif atom_id.startswith('O'):
return 'O'
elif atom_id.startswith('H'):
return 'H'
elif atom_id.startswith('S'):
return 'S'
raise ValueError('Atom ID not recognized.')
# NB: restype_3to1 differs from Bio.PDB.protein_letters_3to1 by being a simple
# 1-to-1 mapping of 3 letter names to one letter names. The latter contains
# many more, and less common, three letter names as keys and maps many of these
# to the same one letter name (including 'X' and 'U' which we don't use here).
restype_3to1 = {v: k for k, v in restype_1to3.items()}
# Define a restype name for all unknown residues.
unk_restype = 'UNK'
resnames = [restype_1to3[r] for r in restypes] + [unk_restype]
resname_to_idx = {resname: i for i, resname in enumerate(resnames)}
# The mapping here uses hhblits convention, so that B is mapped to D, J and O
# are mapped to X, U is mapped to C, and Z is mapped to E. Other than that the
# remaining 20 amino acids are kept in alphabetical order.
# There are 2 non-amino acid codes, X (representing any amino acid) and
# "-" representing a missing amino acid in an alignment. The id for these
# codes is put at the end (20 and 21) so that they can easily be ignored if
# desired.
HHBLITS_AA_TO_ID = {
'A': 0,
'B': 2,
'C': 1,
'D': 2,
'E': 3,
'F': 4,
'G': 5,
'H': 6,
'I': 7,
'J': 20,
'K': 8,
'L': 9,
'M': 10,
'N': 11,
'O': 20,
'P': 12,
'Q': 13,
'R': 14,
'S': 15,
'T': 16,
'U': 1,
'V': 17,
'W': 18,
'X': 20,
'Y': 19,
'Z': 3,
'-': 21,
}
# Partial inversion of HHBLITS_AA_TO_ID.
ID_TO_HHBLITS_AA = {
0: 'A',
1: 'C', # Also U.
2: 'D', # Also B.
3: 'E', # Also Z.
4: 'F',
5: 'G',
6: 'H',
7: 'I',
8: 'K',
9: 'L',
10: 'M',
11: 'N',
12: 'P',
13: 'Q',
14: 'R',
15: 'S',
16: 'T',
17: 'V',
18: 'W',
19: 'Y',
20: 'X', # Includes J and O.
21: '-',
}
restypes_with_x_and_gap = restypes + ['X', '-']
MAP_HHBLITS_AATYPE_TO_OUR_AATYPE = tuple(
restypes_with_x_and_gap.index(ID_TO_HHBLITS_AA[i])
for i in range(len(restypes_with_x_and_gap)))
def _make_standard_atom_mask() -> np.ndarray:
"""Returns [num_res_types, num_atom_types] mask array."""
# +1 to account for unknown (all 0s).
mask = np.zeros([restype_num + 1, atom_type_num], dtype=np.int32)
for restype, restype_letter in enumerate(restypes):
restype_name = restype_1to3[restype_letter]
atom_names = residue_atoms[restype_name]
for atom_name in atom_names:
atom_type = atom_order[atom_name]
mask[restype, atom_type] = 1
return mask
STANDARD_ATOM_MASK = _make_standard_atom_mask()
# A one hot representation for the first and second atoms defining the axis
# of rotation for each chi-angle in each residue.
def chi_angle_atom(atom_index: int) -> np.ndarray:
"""Define chi-angle rigid groups via one-hot representations."""
chi_angles_index = {}
one_hots = []
for k, v in chi_angles_atoms.items():
indices = [atom_types.index(s[atom_index]) for s in v]
indices.extend([-1]*(4-len(indices)))
chi_angles_index[k] = indices
for r in restypes:
res3 = restype_1to3[r]
one_hot = np.eye(atom_type_num)[chi_angles_index[res3]]
one_hots.append(one_hot)
one_hots.append(np.zeros([4, atom_type_num])) # Add zeros for residue `X`.
one_hot = np.stack(one_hots, axis=0)
one_hot = np.transpose(one_hot, [0, 2, 1])
return one_hot
chi_atom_1_one_hot = chi_angle_atom(1)
chi_atom_2_one_hot = chi_angle_atom(2)
# An array like chi_angles_atoms but using indices rather than names.
chi_angles_atom_indices = [chi_angles_atoms[restype_1to3[r]] for r in restypes]
chi_angles_atom_indices = tree.map_structure(
lambda atom_name: atom_order[atom_name], chi_angles_atom_indices)
chi_angles_atom_indices = np.array([
chi_atoms + ([[0, 0, 0, 0]] * (4 - len(chi_atoms)))
for chi_atoms in chi_angles_atom_indices])
# Mapping from (res_name, atom_name) pairs to the atom's chi group index
# and atom index within that group.
chi_groups_for_atom = collections.defaultdict(list)
for res_name, chi_angle_atoms_for_res in chi_angles_atoms.items():
for chi_group_i, chi_group in enumerate(chi_angle_atoms_for_res):
for atom_i, atom in enumerate(chi_group):
chi_groups_for_atom[(res_name, atom)].append((chi_group_i, atom_i))
chi_groups_for_atom = dict(chi_groups_for_atom)
def _make_rigid_transformation_4x4(ex, ey, translation):
"""Create a rigid 4x4 transformation matrix from two axes and transl."""
# Normalize ex.
ex_normalized = ex / np.linalg.norm(ex)
# make ey perpendicular to ex
ey_normalized = ey - np.dot(ey, ex_normalized) * ex_normalized
ey_normalized /= np.linalg.norm(ey_normalized)
# compute ez as cross product
eznorm = np.cross(ex_normalized, ey_normalized)
m = np.stack([ex_normalized, ey_normalized, eznorm, translation]).transpose()
m = np.concatenate([m, [[0., 0., 0., 1.]]], axis=0)
return m
# create an array with (restype, atomtype) --> rigid_group_idx
# and an array with (restype, atomtype, coord) for the atom positions
# and compute affine transformation matrices (4,4) from one rigid group to the
# previous group
restype_atom37_to_rigid_group = np.zeros([21, 37], dtype=int)
restype_atom37_mask = np.zeros([21, 37], dtype=np.float32)
restype_atom37_rigid_group_positions = np.zeros([21, 37, 3], dtype=np.float32)
restype_atom14_to_rigid_group = np.zeros([21, 14], dtype=int)
restype_atom14_mask = np.zeros([21, 14], dtype=np.float32)
restype_atom14_rigid_group_positions = np.zeros([21, 14, 3], dtype=np.float32)
restype_rigid_group_default_frame = np.zeros([21, 8, 4, 4], dtype=np.float32)
def _make_rigid_group_constants():
"""Fill the arrays above."""
for restype, restype_letter in enumerate(restypes):
resname = restype_1to3[restype_letter]
for atomname, group_idx, atom_position in rigid_group_atom_positions[
resname]:
atomtype = atom_order[atomname]
restype_atom37_to_rigid_group[restype, atomtype] = group_idx
restype_atom37_mask[restype, atomtype] = 1
restype_atom37_rigid_group_positions[restype, atomtype, :] = atom_position
atom14idx = restype_name_to_atom14_names[resname].index(atomname)
restype_atom14_to_rigid_group[restype, atom14idx] = group_idx
restype_atom14_mask[restype, atom14idx] = 1
restype_atom14_rigid_group_positions[restype,
atom14idx, :] = atom_position
for restype, restype_letter in enumerate(restypes):
resname = restype_1to3[restype_letter]
atom_positions = {name: np.array(pos) for name, _, pos
in rigid_group_atom_positions[resname]}
# backbone to backbone is the identity transform
restype_rigid_group_default_frame[restype, 0, :, :] = np.eye(4)
# pre-omega-frame to backbone (currently dummy identity matrix)
restype_rigid_group_default_frame[restype, 1, :, :] = np.eye(4)
# phi-frame to backbone
mat = _make_rigid_transformation_4x4(
ex=atom_positions['N'] - atom_positions['CA'],
ey=np.array([1., 0., 0.]),
translation=atom_positions['N'])
restype_rigid_group_default_frame[restype, 2, :, :] = mat
# psi-frame to backbone
mat = _make_rigid_transformation_4x4(
ex=atom_positions['C'] - atom_positions['CA'],
ey=atom_positions['CA'] - atom_positions['N'],
translation=atom_positions['C'])
restype_rigid_group_default_frame[restype, 3, :, :] = mat
# chi1-frame to backbone
if chi_angles_mask[restype][0]:
base_atom_names = chi_angles_atoms[resname][0]
base_atom_positions = [atom_positions[name] for name in base_atom_names]
mat = _make_rigid_transformation_4x4(
ex=base_atom_positions[2] - base_atom_positions[1],
ey=base_atom_positions[0] - base_atom_positions[1],
translation=base_atom_positions[2])
restype_rigid_group_default_frame[restype, 4, :, :] = mat
# chi2-frame to chi1-frame
# chi3-frame to chi2-frame
# chi4-frame to chi3-frame
# luckily all rotation axes for the next frame start at (0,0,0) of the
# previous frame
for chi_idx in range(1, 4):
if chi_angles_mask[restype][chi_idx]:
axis_end_atom_name = chi_angles_atoms[resname][chi_idx][2]
axis_end_atom_position = atom_positions[axis_end_atom_name]
mat = _make_rigid_transformation_4x4(
ex=axis_end_atom_position,
ey=np.array([-1., 0., 0.]),
translation=axis_end_atom_position)
restype_rigid_group_default_frame[restype, 4 + chi_idx, :, :] = mat
_make_rigid_group_constants()
def make_atom14_dists_bounds(overlap_tolerance=1.5,
bond_length_tolerance_factor=15):
"""compute upper and lower bounds for bonds to assess violations."""
restype_atom14_bond_lower_bound = np.zeros([21, 14, 14], np.float32)
restype_atom14_bond_upper_bound = np.zeros([21, 14, 14], np.float32)
restype_atom14_bond_stddev = np.zeros([21, 14, 14], np.float32)
residue_bonds, residue_virtual_bonds, _ = load_stereo_chemical_props()
for restype, restype_letter in enumerate(restypes):
resname = restype_1to3[restype_letter]
atom_list = restype_name_to_atom14_names[resname]
# create lower and upper bounds for clashes
for atom1_idx, atom1_name in enumerate(atom_list):
if not atom1_name:
continue
atom1_radius = van_der_waals_radius[atom1_name[0]]
for atom2_idx, atom2_name in enumerate(atom_list):
if (not atom2_name) or atom1_idx == atom2_idx:
continue
atom2_radius = van_der_waals_radius[atom2_name[0]]
lower = atom1_radius + atom2_radius - overlap_tolerance
upper = 1e10
restype_atom14_bond_lower_bound[restype, atom1_idx, atom2_idx] = lower
restype_atom14_bond_lower_bound[restype, atom2_idx, atom1_idx] = lower
restype_atom14_bond_upper_bound[restype, atom1_idx, atom2_idx] = upper
restype_atom14_bond_upper_bound[restype, atom2_idx, atom1_idx] = upper
# overwrite lower and upper bounds for bonds and angles
for b in residue_bonds[resname] + residue_virtual_bonds[resname]:
atom1_idx = atom_list.index(b.atom1_name)
atom2_idx = atom_list.index(b.atom2_name)
lower = b.length - bond_length_tolerance_factor * b.stddev
upper = b.length + bond_length_tolerance_factor * b.stddev
restype_atom14_bond_lower_bound[restype, atom1_idx, atom2_idx] = lower
restype_atom14_bond_lower_bound[restype, atom2_idx, atom1_idx] = lower
restype_atom14_bond_upper_bound[restype, atom1_idx, atom2_idx] = upper
restype_atom14_bond_upper_bound[restype, atom2_idx, atom1_idx] = upper
restype_atom14_bond_stddev[restype, atom1_idx, atom2_idx] = b.stddev
restype_atom14_bond_stddev[restype, atom2_idx, atom1_idx] = b.stddev
return {'lower_bound': restype_atom14_bond_lower_bound, # shape (21,14,14)
'upper_bound': restype_atom14_bond_upper_bound, # shape (21,14,14)
'stddev': restype_atom14_bond_stddev, # shape (21,14,14)
}
alphafold/common/residue_constants_test.py 0 → 100644
View file @ eb93322b
# Copyright 2021 DeepMind Technologies Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Test that residue_constants generates correct values."""
from absl.testing import absltest
from absl.testing import parameterized
from alphafold.common import residue_constants
import numpy as np
class ResidueConstantsTest(parameterized.TestCase):
@parameterized.parameters(
('ALA', 0),
('CYS', 1),
('HIS', 2),
('MET', 3),
('LYS', 4),
('ARG', 4),
)
def testChiAnglesAtoms(self, residue_name, chi_num):
chi_angles_atoms = residue_constants.chi_angles_atoms[residue_name]
self.assertLen(chi_angles_atoms, chi_num)
for chi_angle_atoms in chi_angles_atoms:
self.assertLen(chi_angle_atoms, 4)
def testChiGroupsForAtom(self):
for k, chi_groups in residue_constants.chi_groups_for_atom.items():
res_name, atom_name = k
for chi_group_i, atom_i in chi_groups:
self.assertEqual(
atom_name,
residue_constants.chi_angles_atoms[res_name][chi_group_i][atom_i])
@parameterized.parameters(
('ALA', 5), ('ARG', 11), ('ASN', 8), ('ASP', 8), ('CYS', 6), ('GLN', 9),
('GLU', 9), ('GLY', 4), ('HIS', 10), ('ILE', 8), ('LEU', 8), ('LYS', 9),
('MET', 8), ('PHE', 11), ('PRO', 7), ('SER', 6), ('THR', 7), ('TRP', 14),
('TYR', 12), ('VAL', 7)
)
def testResidueAtoms(self, atom_name, num_residue_atoms):
residue_atoms = residue_constants.residue_atoms[atom_name]
self.assertLen(residue_atoms, num_residue_atoms)
def testStandardAtomMask(self):
with self.subTest('Check shape'):
self.assertEqual(residue_constants.STANDARD_ATOM_MASK.shape, (21, 37,))
with self.subTest('Check values'):
str_to_row = lambda s: [c == '1' for c in s] # More clear/concise.
np.testing.assert_array_equal(
residue_constants.STANDARD_ATOM_MASK,
np.array([
# NB This was defined by c+p but looks sane.
str_to_row('11111 '), # ALA
str_to_row('111111 1 1 11 1 '), # ARG
str_to_row('111111 11 '), # ASP
str_to_row('111111 11 '), # ASN
str_to_row('11111 1 '), # CYS
str_to_row('111111 1 11 '), # GLU
str_to_row('111111 1 11 '), # GLN
str_to_row('111 1 '), # GLY
str_to_row('111111 11 1 1 '), # HIS
str_to_row('11111 11 1 '), # ILE
str_to_row('111111 11 '), # LEU
str_to_row('111111 1 1 1 '), # LYS
str_to_row('111111 11 '), # MET
str_to_row('111111 11 11 1 '), # PHE
str_to_row('111111 1 '), # PRO
str_to_row('11111 1 '), # SER
str_to_row('11111 1 1 '), # THR
str_to_row('111111 11 11 1 1 11 '), # TRP
str_to_row('111111 11 11 11 '), # TYR
str_to_row('11111 11 '), # VAL
str_to_row(' '), # UNK
]))
with self.subTest('Check row totals'):
# Check each row has the right number of atoms.
for row, restype in enumerate(residue_constants.restypes): # A, R, ...
long_restype = residue_constants.restype_1to3[restype] # ALA, ARG, ...
atoms_names = residue_constants.residue_atoms[
long_restype] # ['C', 'CA', 'CB', 'N', 'O'], ...
self.assertLen(atoms_names,
residue_constants.STANDARD_ATOM_MASK[row, :].sum(),
long_restype)
def testAtomTypes(self):
self.assertEqual(residue_constants.atom_type_num, 37)
self.assertEqual(residue_constants.atom_types[0], 'N')
self.assertEqual(residue_constants.atom_types[1], 'CA')
self.assertEqual(residue_constants.atom_types[2], 'C')
self.assertEqual(residue_constants.atom_types[3], 'CB')
self.assertEqual(residue_constants.atom_types[4], 'O')
self.assertEqual(residue_constants.atom_order['N'], 0)
self.assertEqual(residue_constants.atom_order['CA'], 1)
self.assertEqual(residue_constants.atom_order['C'], 2)
self.assertEqual(residue_constants.atom_order['CB'], 3)
self.assertEqual(residue_constants.atom_order['O'], 4)
self.assertEqual(residue_constants.atom_type_num, 37)
def testRestypes(self):
three_letter_restypes = [
residue_constants.restype_1to3[r] for r in residue_constants.restypes]
for restype, exp_restype in zip(
three_letter_restypes, sorted(residue_constants.restype_1to3.values())):
self.assertEqual(restype, exp_restype)
self.assertEqual(residue_constants.restype_num, 20)
def testSequenceToOneHotHHBlits(self):
one_hot = residue_constants.sequence_to_onehot(
'ABCDEFGHIJKLMNOPQRSTUVWXYZ-', residue_constants.HHBLITS_AA_TO_ID)
exp_one_hot = np.array(
[[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0],
[0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1]])
np.testing.assert_array_equal(one_hot, exp_one_hot)
def testSequenceToOneHotStandard(self):
one_hot = residue_constants.sequence_to_onehot(
'ARNDCQEGHILKMFPSTWYV', residue_constants.restype_order)
np.testing.assert_array_equal(one_hot, np.eye(20))
def testSequenceToOneHotUnknownMapping(self):
seq = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
expected_out = np.zeros([26, 21])
for row, position in enumerate(
[0, 20, 4, 3, 6, 13, 7, 8, 9, 20, 11, 10, 12, 2, 20, 14, 5, 1, 15, 16,
20, 19, 17, 20, 18, 20]):
expected_out[row, position] = 1
aa_types = residue_constants.sequence_to_onehot(
sequence=seq,
mapping=residue_constants.restype_order_with_x,
map_unknown_to_x=True)
self.assertTrue((aa_types == expected_out).all())
@parameterized.named_parameters(
('lowercase', 'aaa'), # Insertions in A3M.
('gaps', '---'), # Gaps in A3M.
('dots', '...'), # Gaps in A3M.
('metadata', '>TEST'), # FASTA metadata line.
)
def testSequenceToOneHotUnknownMappingError(self, seq):
with self.assertRaises(ValueError):
residue_constants.sequence_to_onehot(
sequence=seq,
mapping=residue_constants.restype_order_with_x,
map_unknown_to_x=True)
if __name__ == '__main__':
absltest.main()
alphafold/common/stereo_chemical_props.txt 0 → 100644
View file @ eb93322b
Bond Residue Mean StdDev
CA-CB ALA 1.520 0.021
N-CA ALA 1.459 0.020
CA-C ALA 1.525 0.026
C-O ALA 1.229 0.019
CA-CB ARG 1.535 0.022
CB-CG ARG 1.521 0.027
CG-CD ARG 1.515 0.025
CD-NE ARG 1.460 0.017
NE-CZ ARG 1.326 0.013
CZ-NH1 ARG 1.326 0.013
CZ-NH2 ARG 1.326 0.013
N-CA ARG 1.459 0.020
CA-C ARG 1.525 0.026
C-O ARG 1.229 0.019
CA-CB ASN 1.527 0.026
CB-CG ASN 1.506 0.023
CG-OD1 ASN 1.235 0.022
CG-ND2 ASN 1.324 0.025
N-CA ASN 1.459 0.020
CA-C ASN 1.525 0.026
C-O ASN 1.229 0.019
CA-CB ASP 1.535 0.022
CB-CG ASP 1.513 0.021
CG-OD1 ASP 1.249 0.023
CG-OD2 ASP 1.249 0.023
N-CA ASP 1.459 0.020
CA-C ASP 1.525 0.026
C-O ASP 1.229 0.019
CA-CB CYS 1.526 0.013
CB-SG CYS 1.812 0.016
N-CA CYS 1.459 0.020
CA-C CYS 1.525 0.026
C-O CYS 1.229 0.019
CA-CB GLU 1.535 0.022
CB-CG GLU 1.517 0.019
CG-CD GLU 1.515 0.015
CD-OE1 GLU 1.252 0.011
CD-OE2 GLU 1.252 0.011
N-CA GLU 1.459 0.020
CA-C GLU 1.525 0.026
C-O GLU 1.229 0.019
CA-CB GLN 1.535 0.022
CB-CG GLN 1.521 0.027
CG-CD GLN 1.506 0.023
CD-OE1 GLN 1.235 0.022
CD-NE2 GLN 1.324 0.025
N-CA GLN 1.459 0.020
CA-C GLN 1.525 0.026
C-O GLN 1.229 0.019
N-CA GLY 1.456 0.015
CA-C GLY 1.514 0.016
C-O GLY 1.232 0.016
CA-CB HIS 1.535 0.022
CB-CG HIS 1.492 0.016
CG-ND1 HIS 1.369 0.015
CG-CD2 HIS 1.353 0.017
ND1-CE1 HIS 1.343 0.025
CD2-NE2 HIS 1.415 0.021
CE1-NE2 HIS 1.322 0.023
N-CA HIS 1.459 0.020
CA-C HIS 1.525 0.026
C-O HIS 1.229 0.019
CA-CB ILE 1.544 0.023
CB-CG1 ILE 1.536 0.028
CB-CG2 ILE 1.524 0.031
CG1-CD1 ILE 1.500 0.069
N-CA ILE 1.459 0.020
CA-C ILE 1.525 0.026
C-O ILE 1.229 0.019
CA-CB LEU 1.533 0.023
CB-CG LEU 1.521 0.029
CG-CD1 LEU 1.514 0.037
CG-CD2 LEU 1.514 0.037
N-CA LEU 1.459 0.020
CA-C LEU 1.525 0.026
C-O LEU 1.229 0.019
CA-CB LYS 1.535 0.022
CB-CG LYS 1.521 0.027
CG-CD LYS 1.520 0.034
CD-CE LYS 1.508 0.025
CE-NZ LYS 1.486 0.025
N-CA LYS 1.459 0.020
CA-C LYS 1.525 0.026
C-O LYS 1.229 0.019
CA-CB MET 1.535 0.022
CB-CG MET 1.509 0.032
CG-SD MET 1.807 0.026
SD-CE MET 1.774 0.056
N-CA MET 1.459 0.020
CA-C MET 1.525 0.026
C-O MET 1.229 0.019
CA-CB PHE 1.535 0.022
CB-CG PHE 1.509 0.017
CG-CD1 PHE 1.383 0.015
CG-CD2 PHE 1.383 0.015
CD1-CE1 PHE 1.388 0.020
CD2-CE2 PHE 1.388 0.020
CE1-CZ PHE 1.369 0.019
CE2-CZ PHE 1.369 0.019
N-CA PHE 1.459 0.020
CA-C PHE 1.525 0.026
C-O PHE 1.229 0.019
CA-CB PRO 1.531 0.020
CB-CG PRO 1.495 0.050
CG-CD PRO 1.502 0.033
CD-N PRO 1.474 0.014
N-CA PRO 1.468 0.017
CA-C PRO 1.524 0.020
C-O PRO 1.228 0.020
CA-CB SER 1.525 0.015
CB-OG SER 1.418 0.013
N-CA SER 1.459 0.020
CA-C SER 1.525 0.026
C-O SER 1.229 0.019
CA-CB THR 1.529 0.026
CB-OG1 THR 1.428 0.020
CB-CG2 THR 1.519 0.033
N-CA THR 1.459 0.020
CA-C THR 1.525 0.026
C-O THR 1.229 0.019
CA-CB TRP 1.535 0.022
CB-CG TRP 1.498 0.018
CG-CD1 TRP 1.363 0.014
CG-CD2 TRP 1.432 0.017
CD1-NE1 TRP 1.375 0.017
NE1-CE2 TRP 1.371 0.013
CD2-CE2 TRP 1.409 0.012
CD2-CE3 TRP 1.399 0.015
CE2-CZ2 TRP 1.393 0.017
CE3-CZ3 TRP 1.380 0.017
CZ2-CH2 TRP 1.369 0.019
CZ3-CH2 TRP 1.396 0.016
N-CA TRP 1.459 0.020
CA-C TRP 1.525 0.026
C-O TRP 1.229 0.019
CA-CB TYR 1.535 0.022
CB-CG TYR 1.512 0.015
CG-CD1 TYR 1.387 0.013
CG-CD2 TYR 1.387 0.013
CD1-CE1 TYR 1.389 0.015
CD2-CE2 TYR 1.389 0.015
CE1-CZ TYR 1.381 0.013
CE2-CZ TYR 1.381 0.013
CZ-OH TYR 1.374 0.017
N-CA TYR 1.459 0.020
CA-C TYR 1.525 0.026
C-O TYR 1.229 0.019
CA-CB VAL 1.543 0.021
CB-CG1 VAL 1.524 0.021
CB-CG2 VAL 1.524 0.021
N-CA VAL 1.459 0.020
CA-C VAL 1.525 0.026
C-O VAL 1.229 0.019
-
Angle Residue Mean StdDev
N-CA-CB ALA 110.1 1.4
CB-CA-C ALA 110.1 1.5
N-CA-C ALA 111.0 2.7
CA-C-O ALA 120.1 2.1
N-CA-CB ARG 110.6 1.8
CB-CA-C ARG 110.4 2.0
CA-CB-CG ARG 113.4 2.2
CB-CG-CD ARG 111.6 2.6
CG-CD-NE ARG 111.8 2.1
CD-NE-CZ ARG 123.6 1.4
NE-CZ-NH1 ARG 120.3 0.5
NE-CZ-NH2 ARG 120.3 0.5
NH1-CZ-NH2 ARG 119.4 1.1
N-CA-C ARG 111.0 2.7
CA-C-O ARG 120.1 2.1
N-CA-CB ASN 110.6 1.8
CB-CA-C ASN 110.4 2.0
CA-CB-CG ASN 113.4 2.2
CB-CG-ND2 ASN 116.7 2.4
CB-CG-OD1 ASN 121.6 2.0
ND2-CG-OD1 ASN 121.9 2.3
N-CA-C ASN 111.0 2.7
CA-C-O ASN 120.1 2.1
N-CA-CB ASP 110.6 1.8
CB-CA-C ASP 110.4 2.0
CA-CB-CG ASP 113.4 2.2
CB-CG-OD1 ASP 118.3 0.9
CB-CG-OD2 ASP 118.3 0.9
OD1-CG-OD2 ASP 123.3 1.9
N-CA-C ASP 111.0 2.7
CA-C-O ASP 120.1 2.1
N-CA-CB CYS 110.8 1.5
CB-CA-C CYS 111.5 1.2
CA-CB-SG CYS 114.2 1.1
N-CA-C CYS 111.0 2.7
CA-C-O CYS 120.1 2.1
N-CA-CB GLU 110.6 1.8
CB-CA-C GLU 110.4 2.0
CA-CB-CG GLU 113.4 2.2
CB-CG-CD GLU 114.2 2.7
CG-CD-OE1 GLU 118.3 2.0
CG-CD-OE2 GLU 118.3 2.0
OE1-CD-OE2 GLU 123.3 1.2
N-CA-C GLU 111.0 2.7
CA-C-O GLU 120.1 2.1
N-CA-CB GLN 110.6 1.8
CB-CA-C GLN 110.4 2.0
CA-CB-CG GLN 113.4 2.2
CB-CG-CD GLN 111.6 2.6
CG-CD-OE1 GLN 121.6 2.0
CG-CD-NE2 GLN 116.7 2.4
OE1-CD-NE2 GLN 121.9 2.3
N-CA-C GLN 111.0 2.7
CA-C-O GLN 120.1 2.1
N-CA-C GLY 113.1 2.5
CA-C-O GLY 120.6 1.8
N-CA-CB HIS 110.6 1.8
CB-CA-C HIS 110.4 2.0
CA-CB-CG HIS 113.6 1.7
CB-CG-ND1 HIS 123.2 2.5
CB-CG-CD2 HIS 130.8 3.1
CG-ND1-CE1 HIS 108.2 1.4
ND1-CE1-NE2 HIS 109.9 2.2
CE1-NE2-CD2 HIS 106.6 2.5
NE2-CD2-CG HIS 109.2 1.9
CD2-CG-ND1 HIS 106.0 1.4
N-CA-C HIS 111.0 2.7
CA-C-O HIS 120.1 2.1
N-CA-CB ILE 110.8 2.3
CB-CA-C ILE 111.6 2.0
CA-CB-CG1 ILE 111.0 1.9
CB-CG1-CD1 ILE 113.9 2.8
CA-CB-CG2 ILE 110.9 2.0
CG1-CB-CG2 ILE 111.4 2.2
N-CA-C ILE 111.0 2.7
CA-C-O ILE 120.1 2.1
N-CA-CB LEU 110.4 2.0
CB-CA-C LEU 110.2 1.9
CA-CB-CG LEU 115.3 2.3
CB-CG-CD1 LEU 111.0 1.7
CB-CG-CD2 LEU 111.0 1.7
CD1-CG-CD2 LEU 110.5 3.0
N-CA-C LEU 111.0 2.7
CA-C-O LEU 120.1 2.1
N-CA-CB LYS 110.6 1.8
CB-CA-C LYS 110.4 2.0
CA-CB-CG LYS 113.4 2.2
CB-CG-CD LYS 111.6 2.6
CG-CD-CE LYS 111.9 3.0
CD-CE-NZ LYS 111.7 2.3
N-CA-C LYS 111.0 2.7
CA-C-O LYS 120.1 2.1
N-CA-CB MET 110.6 1.8
CB-CA-C MET 110.4 2.0
CA-CB-CG MET 113.3 1.7
CB-CG-SD MET 112.4 3.0
CG-SD-CE MET 100.2 1.6
N-CA-C MET 111.0 2.7
CA-C-O MET 120.1 2.1
N-CA-CB PHE 110.6 1.8
CB-CA-C PHE 110.4 2.0
CA-CB-CG PHE 113.9 2.4
CB-CG-CD1 PHE 120.8 0.7
CB-CG-CD2 PHE 120.8 0.7
CD1-CG-CD2 PHE 118.3 1.3
CG-CD1-CE1 PHE 120.8 1.1
CG-CD2-CE2 PHE 120.8 1.1
CD1-CE1-CZ PHE 120.1 1.2
CD2-CE2-CZ PHE 120.1 1.2
CE1-CZ-CE2 PHE 120.0 1.8
N-CA-C PHE 111.0 2.7
CA-C-O PHE 120.1 2.1
N-CA-CB PRO 103.3 1.2
CB-CA-C PRO 111.7 2.1
CA-CB-CG PRO 104.8 1.9
CB-CG-CD PRO 106.5 3.9
CG-CD-N PRO 103.2 1.5
CA-N-CD PRO 111.7 1.4
N-CA-C PRO 112.1 2.6
CA-C-O PRO 120.2 2.4
N-CA-CB SER 110.5 1.5
CB-CA-C SER 110.1 1.9
CA-CB-OG SER 111.2 2.7
N-CA-C SER 111.0 2.7
CA-C-O SER 120.1 2.1
N-CA-CB THR 110.3 1.9
CB-CA-C THR 111.6 2.7
CA-CB-OG1 THR 109.0 2.1
CA-CB-CG2 THR 112.4 1.4
OG1-CB-CG2 THR 110.0 2.3
N-CA-C THR 111.0 2.7
CA-C-O THR 120.1 2.1
N-CA-CB TRP 110.6 1.8
CB-CA-C TRP 110.4 2.0
CA-CB-CG TRP 113.7 1.9
CB-CG-CD1 TRP 127.0 1.3
CB-CG-CD2 TRP 126.6 1.3
CD1-CG-CD2 TRP 106.3 0.8
CG-CD1-NE1 TRP 110.1 1.0
CD1-NE1-CE2 TRP 109.0 0.9
NE1-CE2-CD2 TRP 107.3 1.0
CE2-CD2-CG TRP 107.3 0.8
CG-CD2-CE3 TRP 133.9 0.9
NE1-CE2-CZ2 TRP 130.4 1.1
CE3-CD2-CE2 TRP 118.7 1.2
CD2-CE2-CZ2 TRP 122.3 1.2
CE2-CZ2-CH2 TRP 117.4 1.0
CZ2-CH2-CZ3 TRP 121.6 1.2
CH2-CZ3-CE3 TRP 121.2 1.1
CZ3-CE3-CD2 TRP 118.8 1.3
N-CA-C TRP 111.0 2.7
CA-C-O TRP 120.1 2.1
N-CA-CB TYR 110.6 1.8
CB-CA-C TYR 110.4 2.0
CA-CB-CG TYR 113.4 1.9
CB-CG-CD1 TYR 121.0 0.6
CB-CG-CD2 TYR 121.0 0.6
CD1-CG-CD2 TYR 117.9 1.1
CG-CD1-CE1 TYR 121.3 0.8
CG-CD2-CE2 TYR 121.3 0.8
CD1-CE1-CZ TYR 119.8 0.9
CD2-CE2-CZ TYR 119.8 0.9
CE1-CZ-CE2 TYR 119.8 1.6
CE1-CZ-OH TYR 120.1 2.7
CE2-CZ-OH TYR 120.1 2.7
N-CA-C TYR 111.0 2.7
CA-C-O TYR 120.1 2.1
N-CA-CB VAL 111.5 2.2
CB-CA-C VAL 111.4 1.9
CA-CB-CG1 VAL 110.9 1.5
CA-CB-CG2 VAL 110.9 1.5
CG1-CB-CG2 VAL 110.9 1.6
N-CA-C VAL 111.0 2.7
CA-C-O VAL 120.1 2.1
-
Non-bonded distance Minimum Dist Tolerance
C-C 3.4 1.5
C-N 3.25 1.5
C-S 3.5 1.5
C-O 3.22 1.5
N-N 3.1 1.5
N-S 3.35 1.5
N-O 3.07 1.5
O-S 3.32 1.5
O-O 3.04 1.5
S-S 2.03 1.0
-
alphafold/common/testdata/2rbg.pdb 0 → 100755
View file @ eb93322b
This source diff could not be displayed because it is too large. You can view the blob instead.
alphafold/common/testdata/5nmu.pdb 0 → 100644
View file @ eb93322b
This source diff could not be displayed because it is too large. You can view the blob instead.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment