Skip to content

GitLab

Commit 818f108a authored by zhuwenwen's avatar zhuwenwen
Browse files

update to v2.3.2

parent f7451744
Hide whitespace changes
Inline Side-by-side
Showing with 541 additions and 786 deletions
README.md
View file @ 818f108a
![header](imgs/header.jpg)
<!--
* @Author: zhuww
* @email: zhuww@sugon.com
* @Date: 2023-04-06 18:04:07
* @LastEditTime: 2023-11-15 17:30:01
-->
# AF2
## 论文
- [https://www.nature.com/articles/s41586-021-03819-2](https://www.nature.com/articles/s41586-021-03819-2)
# AlphaFold
## 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。
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.
![img](./docs/alphafold2.png)
We also provide:
## 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。
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.
![img](./docs/alphafold2_1.png)
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).
## 环境配置
提供[光源](https://www.sourcefind.cn/#/service-details)拉取推理的docker镜像：
```
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:alphafold2-2.3.2-dtk-23.10-py38
# <Image ID>用上面拉取docker镜像的ID替换
# <Host Path>主机端路径
# <Container Path>容器映射路径
docker run -it --name alphafold --shm-size=32G --device=/dev/kfd --device=/dev/dri/ --cap-add=SYS_PTRACE --security-opt seccomp=unconfined --ulimit memlock=-1:-1 --ipc=host --network host --group-add video -v <Host Path>:<Container Path> <Image ID> /bin/bash
```
: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.**
镜像版本依赖：
* DTK驱动：dtk23.10
* Jax: 0.3.25
* TensorFlow2: 2.11.0
* python: python3.8
: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.**
激活镜像环境：
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:
`source /opt/dtk-23.10/env.sh`
## 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/ # Total: ~ 2.62 TB (download: 556 GB)
bfd/ # ~ 1.8 TB (download: 271.6 GB)
# 6 files.
mgnify/ # ~ 120 GB (download: 67 GB)
$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/ # ~ 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)
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/
# About 199,000 .cif files.
100d.cif
101d.cif
101m.cif
...
obsolete.dat
pdb_seqres/ # ~ 0.2 GB (download: 0.2 GB)
pdb_seqres/
pdb_seqres.txt
small_bfd/ # ~ 17 GB (download: 9.6 GB)
small_bfd/
bfd-first_non_consensus_sequences.fasta
uniref30/ # ~ 206 GB (download: 52.5 GB)
# 7 files.
uniprot/ # ~ 105 GB (download: 53 GB)
uniref30/
UniRef30_2021_03_hhm.ffindex
UniRef30_2021_03_hhm.ffdata
UniRef30_2021_03_cs219.ffindex
...
uniprot/
uniprot.fasta
uniref90/ # ~ 67 GB (download: 34 GB)
uniref90/
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
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：
**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`:
./scripts/download_all_data.sh 数据集下载目录
## 推理
分别提供了基于Jax的单体和多体的推理脚本.
```bash
python3 docker/run_docker.py \
--fasta_paths=multimer.fasta \
--max_template_date=2020-05-14 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
git clone http://developer.hpccube.com/codes/modelzoo/alphafold2_jax.git # 选择需要的分支下载
cd alphafold2_jax
```
By default the multimer system will run 5 seeds per model (25 total predictions)
for a small drop in accuracy you may wish to run a single seed per model. This
can be done via the `--num_multimer_predictions_per_model` flag, e.g. set it to
`--num_multimer_predictions_per_model=1` to run a single seed per model.
### AlphaFold prediction speed
The table below reports prediction runtimes for proteins of various lengths. We
only measure unrelaxed structure prediction with three recycles while
excluding runtimes from MSA and template search. When running
`docker/run_docker.py` with `--benchmark=true`, this runtime is stored in
`timings.json`. All runtimes are from a single A100 NVIDIA GPU. Prediction
speed on A100 for smaller structures can be improved by increasing
`global_config.subbatch_size` in `alphafold/model/config.py`.
No. residues | Prediction time (s)
-----------: | ------------------:
100 | 4.9
200 | 7.7
300 | 13
400 | 18
500 | 29
600 | 36
700 | 53
800 | 60
900 | 91
1,000 | 96
1,100 | 140
1,500 | 280
2,000 | 450
2,500 | 969
3,000 | 1,240
3,500 | 2,465
4,000 | 5,660
4,500 | 12,475
5,000 | 18,824
### Examples
Below are examples on how to use AlphaFold in different scenarios.
#### Folding a monomer
Say we have a monomer with the sequence `<SEQUENCE>`. The input fasta should be:
```fasta
>sequence_name
<SEQUENCE>
```
Then run the following command:
### 单体
```bash
python3 docker/run_docker.py \
--fasta_paths=monomer.fasta \
--max_template_date=2021-11-01 \
--model_preset=monomer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
./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则可以加载已经搜索对齐的序列，否则默认进行搜索对齐。
#### 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:
### 多体
```bash
python3 docker/run_docker.py \
--fasta_paths=homomer.fasta \
--max_template_date=2021-11-01 \
--model_preset=multimer \
--data_dir=$DOWNLOAD_DIR \
--output_dir=/home/user/absolute_path_to_the_output_dir
./run_multimer.sh
```
多体推理参数说明：multimer.fasta为推理的多体序列，`--model_preset=multimer`为多体模型配置；`--num_multimer_predictions_per_model`为每个模型预测数量，其他参数同单体推理参数说明一致。
#### Folding a heteromer
Say we have an A2B3 heteromer, i.e. with 2 copies of `<SEQUENCE A>` and 3 copies
of `<SEQUENCE B>`. The input fasta should be:
```fasta
>sequence_1
<SEQUENCE A>
>sequence_2
<SEQUENCE A>
>sequence_3
<SEQUENCE B>
>sequence_4
<SEQUENCE B>
>sequence_5
<SEQUENCE B>
```
Then run the following command:
```bash
python3 docker/run_docker.py \
--fasta_paths=heteromer.fasta \
--max_template_date=2021-11-01 \
--model_preset=multimer \
--data_dir=$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:
## result
`--output_dir`目录结构如下：
```
<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
bfd_uniclust_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
查看蛋白质3D结构：[https://www.pdbus.org/3d-view](https://www.pdbus.org/3d-view)
![img](./docs/result_pdb.png)
### Third-party software
## 精度
测试数据：[casp14](https://www.predictioncenter.org/casp14/targetlist.cgi)、[uniprot](https://www.uniprot.org/)，
使用的加速卡:1张 Z100L-32G
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.
1、lddt
见<target_name>/ranking_debug.json中的`plddts`
### Mirrored Databases
2、其它精度值计算：[https://zhanggroup.org/TM-score/](https://zhanggroup.org/TM-score/)
The following databases have been mirrored by DeepMind, and are available with
reference to the following:
准确性数据：
| 数据类型 | 序列类型 | 序列标签 | 序列长度 | GDT-TS | GDT-HA | LDDT | TM score | MaxSub | RMSD |
| :------: | :------: | :------: | :------: |:------: |:------: | :------: | :------: | :------: |:------: |
| fp32 | 单体 | T1026 | 172 | 0.849 | 0.658 | 75.050 | 0.901 | 0.851 | 1.6 |
| fp32 | 单体 | T1053 | 580 | 0.941 | 0.789 | 92.316 | 0.985 | 0.935 | 1.1 |
| fp32 | 单体 | T1091 | 863 | 0.492 | 0.332 | 85.083 | 0.740 | 0.388 | 6.7 |
* [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/).
## 应用场景
* [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.
### 算法类别
NLP
* [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/).
### 热点应用行业
医疗,科研,教育
* [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/).
## 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)
\ No newline at end of file
alphafold/common/stereo_chemical_props.txt 0 → 100644
View file @ 818f108a
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/relax/amber_minimize.py
View file @ 818f108a
......@@ -27,10 +27,19 @@ from alphafold.relax import utils
import ml_collections
import numpy as np
import jax
from simtk import openmm
from simtk import unit
from simtk.openmm import app as openmm_app
from simtk.openmm.app.internal.pdbstructure import PdbStructure
try:
# openmm >= 7.6
import openmm
from openmm import unit
from openmm import app as openmm_app
from openmm.app.internal.pdbstructure import PdbStructure
except ImportError:
# openmm < 7.6
from simtk import openmm
from simtk import unit
from simtk.openmm import app as openmm_app
from simtk.openmm.app.internal.pdbstructure import PdbStructure
ENERGY = unit.kilocalories_per_mole
......@@ -92,7 +101,7 @@ def _openmm_minimize(
_add_restraints(system, pdb, stiffness, restraint_set, exclude_residues)
integrator = openmm.LangevinIntegrator(0, 0.01, 0.0)
platform = openmm.Platform.getPlatformByName("CUDA" if use_gpu else "CPU")
platform = openmm.Platform.getPlatformByName("HIP" if use_gpu else "CPU")
simulation = openmm_app.Simulation(
pdb.topology, system, integrator, platform)
simulation.context.setPositions(pdb.positions)
......
requirements.txt
View file @ 818f108a
......@@ -3,11 +3,11 @@ biopython==1.79
chex==0.0.7
dm-haiku==0.0.9
dm-tree==0.1.6
docker==5.0.0
# docker==5.0.0
immutabledict==2.0.0
jax==0.3.25
# jax==0.3.25
ml-collections==0.1.0
numpy==1.21.6
pandas==1.3.4
scipy==1.7.0
tensorflow-cpu==2.11.0
# tensorflow-cpu==2.11.0
run_alphafold.py
View file @ 818f108a
......@@ -60,6 +60,7 @@ flags.DEFINE_list(
'basename is used to name the output directories for each prediction.')
flags.DEFINE_string('data_dir', None, 'Path to directory of supporting data.')
flags.DEFINE_string('model_names', None, 'Names of models to use.')
flags.DEFINE_string('output_dir', None, 'Path to a directory that will '
'store the results.')
flags.DEFINE_string('jackhmmer_binary_path', shutil.which('jackhmmer'),
......@@ -408,7 +409,8 @@ def main(argv):
data_pipeline = monomer_data_pipeline
model_runners = {}
model_names = config.MODEL_PRESETS[FLAGS.model_preset]
# model_names = config.MODEL_PRESETS[FLAGS.model_preset]
model_names = FLAGS.model_names
for model_name in model_names:
model_config = config.model_config(model_name)
if run_multimer_system:
......@@ -457,6 +459,7 @@ if __name__ == '__main__':
'fasta_paths',
'output_dir',
'data_dir',
'model_names',
'uniref90_database_path',
'mgnify_database_path',
'template_mmcif_dir',
......
run_monomer.sh 0 → 100644
View file @ 818f108a
download_dir=/data/alphafold2
python3 run_alphafold.py \
--fasta_paths=monomer.fasta \
--output_dir=./ \
--use_precomputed_msas=false \
--data_dir=$download_dir \
--model_names="model_1" \
--uniref90_database_path=$download_dir/uniref90/uniref90.fasta \
--mgnify_database_path=$download_dir/mgnify/mgy_clusters_2022_05.fa \
--bfd_database_path=$download_dir/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
--uniref30_database_path=$download_dir/uniref30/UniRef30_2021_03 \
--pdb70_database_path=$download_dir/pdb70/pdb70 \
--template_mmcif_dir=$download_dir/pdb_mmcif/mmcif_files \
--obsolete_pdbs_path=$download_dir/pdb_mmcif/obsolete.dat \
--max_template_date=2020-05-14 \
--model_preset=monomer \
--db_preset=full_dbs \
--models_to_relax=best \
--use_gpu_relax=false
run_multimer.sh 0 → 100644
View file @ 818f108a
download_dir=/data/alphafold2
python3 run_alphafold.py \
--fasta_paths=multimer.fasta \
--output_dir=./ \
--num_multimer_predictions_per_model=1 \
--use_precomputed_msas=false \
--data_dir=$download_dir \
--model_names="model_1_multimer_v3" \
--uniref90_database_path=$download_dir/uniref90/uniref90.fasta \
--mgnify_database_path=$download_dir/mgnify/mgy_clusters_2022_05.fa \
--bfd_database_path=$download_dir/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
--uniref30_database_path=$download_dir/uniref30/UniRef30_2021_03 \
--uniprot_database_path=$download_dir/uniprot/uniprot.fasta \
--pdb_seqres_database_path=$download_dir/pdb_seqres/pdb_seqres.txt \
--template_mmcif_dir=$download_dir/pdb_mmcif/mmcif_files \
--obsolete_pdbs_path=$download_dir/pdb_mmcif/obsolete.dat \
--max_template_date=2020-05-14 \
--model_preset=multimer \
--db_preset=full_dbs \
--models_to_relax=best \
--use_gpu_relax=false
scripts/download_mgnify.sh
View file @ 818f108a
......@@ -38,6 +38,6 @@ BASENAME=$(basename "${SOURCE_URL}")
mkdir --parents "${ROOT_DIR}"
aria2c "${SOURCE_URL}" --dir="${ROOT_DIR}"
pushd "${ROOT_DIR}"
# pushd "${ROOT_DIR}"
gunzip "${ROOT_DIR}/${BASENAME}"
popd
# popd
scripts/download_pdb_mmcif.sh
View file @ 818f108a
......@@ -45,9 +45,24 @@ echo " * rsync.ebi.ac.uk::pub/databases/pdb/data/structures/divided/mmCIF/ (Eur
echo " * ftp.pdbj.org::ftp_data/structures/divided/mmCIF/ (Asia)"
echo "or see https://www.wwpdb.org/ftp/pdb-ftp-sites for more download options."
mkdir --parents "${RAW_DIR}"
rsync --recursive --links --perms --times --compress --info=progress2 --delete --port=33444 \
rsync.rcsb.org::ftp_data/structures/divided/mmCIF/ \
"${RAW_DIR}"
# rsync --recursive --links --perms --times --compress --info=progress2 --delete --port=33444 \
# rsync.rcsb.org::ftp_data/structures/divided/mmCIF/ \
# "${RAW_DIR}"
# (Asia)
rsync -rlpt -v -z --info=progress2 --delete \
ftp.pdbj.org::ftp_data/structures/divided/mmCIF/ \
"${RAW_DIR}"
# (Europe)
# rsync -rlpt -v -z --info=progress2 --delete \
# rsync.ebi.ac.uk::pub/databases/pdb/data/structures/divided/mmCIF/ \
# "${RAW_DIR}"
# fast
# rsync --recursive --links --perms --times --compress --info=progress2 --delete \
# data.pdbj.org::ftp_data/structures/divided/mmCIF/ "${RAW_DIR}"
echo "Unzipping all mmCIF files..."
find "${RAW_DIR}/" -type f -iname "*.gz" -exec gunzip {} +
......
scripts/download_small_bfd.sh
View file @ 818f108a
......@@ -36,6 +36,6 @@ BASENAME=$(basename "${SOURCE_URL}")
mkdir --parents "${ROOT_DIR}"
aria2c "${SOURCE_URL}" --dir="${ROOT_DIR}"
pushd "${ROOT_DIR}"
# pushd "${ROOT_DIR}"
gunzip "${ROOT_DIR}/${BASENAME}"
popd
# popd
scripts/download_uniprot.sh
View file @ 818f108a
......@@ -44,7 +44,7 @@ SPROT_UNZIPPED_BASENAME="${SPROT_BASENAME%.gz}"
mkdir --parents "${ROOT_DIR}"
aria2c "${TREMBL_SOURCE_URL}" --dir="${ROOT_DIR}"
aria2c "${SPROT_SOURCE_URL}" --dir="${ROOT_DIR}"
pushd "${ROOT_DIR}"
# pushd "${ROOT_DIR}"
gunzip "${ROOT_DIR}/${TREMBL_BASENAME}"
gunzip "${ROOT_DIR}/${SPROT_BASENAME}"
......@@ -52,4 +52,4 @@ gunzip "${ROOT_DIR}/${SPROT_BASENAME}"
cat "${ROOT_DIR}/${SPROT_UNZIPPED_BASENAME}" >> "${ROOT_DIR}/${TREMBL_UNZIPPED_BASENAME}"
mv "${ROOT_DIR}/${TREMBL_UNZIPPED_BASENAME}" "${ROOT_DIR}/uniprot.fasta"
rm "${ROOT_DIR}/${SPROT_UNZIPPED_BASENAME}"
popd
# popd
scripts/download_uniref90.sh
View file @ 818f108a
......@@ -36,6 +36,6 @@ BASENAME=$(basename "${SOURCE_URL}")
mkdir --parents "${ROOT_DIR}"
aria2c "${SOURCE_URL}" --dir="${ROOT_DIR}"
pushd "${ROOT_DIR}"
# pushd "${ROOT_DIR}"
gunzip "${ROOT_DIR}/${BASENAME}"
popd
# popd
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