Added switch in the MMCIF processing pipeline for using sequence embeddings when in `seqemb` mode.

- Use sequence embedding files when in `seqemb` mode.
- Make dummy MSA features for MMCIF when using `seqemb` mode.

openfold/data/data_pipeline.py

@@ -702,6 +702,7 @@ class DataPipeline:
         alignment_dir: str,
         chain_id: Optional[str] = None,
         alignment_index: Optional[str] = None,
+        seqemb_mode: bool = False,
     ) -> FeatureDict:
         """
             Assembles features for a specific chain in an mmCIF object.
@@ -727,9 +728,14 @@ class DataPipeline:
             query_release_date=to_date(mmcif.header["release_date"])
         )
 
+        sequence_embedding_features = {}
+        if seqemb_mode:
+            msa_features = make_dummy_msa_feats(input_sequence)
+            sequence_embedding_features = self._process_seqemb_features(alignment_dir)
+        else:
             msa_features = self._process_msa_feats(alignment_dir, input_sequence, alignment_index)
 
-        return {**mmcif_feats, **template_features, **msa_features}
+        return {**mmcif_feats, **template_features, **msa_features, **sequence_embedding_features}
 
     def process_pdb(
         self,