Protein language model-embedded geometric graphs power inter-protein contact prediction
Accurate prediction of contacting residue pairs between interacting proteins is very useful for structural characterization of protein–protein interactions. Although significant improvement has been made in inter-protein contact prediction recently, there is still a large room for improving the pred...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
eLife Sciences Publications Ltd
2024-04-01
|
| Series: | eLife |
| Subjects: | |
| Online Access: | https://elifesciences.org/articles/92184 |
| _version_ | 1797225667417341952 |
|---|---|
| author | Yunda Si Chengfei Yan |
| author_facet | Yunda Si Chengfei Yan |
| author_sort | Yunda Si |
| collection | DOAJ |
| description | Accurate prediction of contacting residue pairs between interacting proteins is very useful for structural characterization of protein–protein interactions. Although significant improvement has been made in inter-protein contact prediction recently, there is still a large room for improving the prediction accuracy. Here we present a new deep learning method referred to as PLMGraph-Inter for inter-protein contact prediction. Specifically, we employ rotationally and translationally invariant geometric graphs obtained from structures of interacting proteins to integrate multiple protein language models, which are successively transformed by graph encoders formed by geometric vector perceptrons and residual networks formed by dimensional hybrid residual blocks to predict inter-protein contacts. Extensive evaluation on multiple test sets illustrates that PLMGraph-Inter outperforms five top inter-protein contact prediction methods, including DeepHomo, GLINTER, CDPred, DeepHomo2, and DRN-1D2D_Inter, by large margins. In addition, we also show that the prediction of PLMGraph-Inter can complement the result of AlphaFold-Multimer. Finally, we show leveraging the contacts predicted by PLMGraph-Inter as constraints for protein–protein docking can dramatically improve its performance for protein complex structure prediction. |
| first_indexed | 2024-04-24T14:12:39Z |
| format | Article |
| id | doaj.art-20faab102b134ac0a45c5b858a91fa76 |
| institution | Directory Open Access Journal |
| issn | 2050-084X |
| language | English |
| last_indexed | 2024-04-24T14:12:39Z |
| publishDate | 2024-04-01 |
| publisher | eLife Sciences Publications Ltd |
| record_format | Article |
| series | eLife |
| spelling | doaj.art-20faab102b134ac0a45c5b858a91fa762024-04-03T08:59:21ZengeLife Sciences Publications LtdeLife2050-084X2024-04-011210.7554/eLife.92184Protein language model-embedded geometric graphs power inter-protein contact predictionYunda Si0Chengfei Yan1https://orcid.org/0000-0002-2010-6668School of Physics, Huazhong University of Science and Technology, Wuhan, ChinaSchool of Physics, Huazhong University of Science and Technology, Wuhan, ChinaAccurate prediction of contacting residue pairs between interacting proteins is very useful for structural characterization of protein–protein interactions. Although significant improvement has been made in inter-protein contact prediction recently, there is still a large room for improving the prediction accuracy. Here we present a new deep learning method referred to as PLMGraph-Inter for inter-protein contact prediction. Specifically, we employ rotationally and translationally invariant geometric graphs obtained from structures of interacting proteins to integrate multiple protein language models, which are successively transformed by graph encoders formed by geometric vector perceptrons and residual networks formed by dimensional hybrid residual blocks to predict inter-protein contacts. Extensive evaluation on multiple test sets illustrates that PLMGraph-Inter outperforms five top inter-protein contact prediction methods, including DeepHomo, GLINTER, CDPred, DeepHomo2, and DRN-1D2D_Inter, by large margins. In addition, we also show that the prediction of PLMGraph-Inter can complement the result of AlphaFold-Multimer. Finally, we show leveraging the contacts predicted by PLMGraph-Inter as constraints for protein–protein docking can dramatically improve its performance for protein complex structure prediction.https://elifesciences.org/articles/92184protein–protein interactioninter-protein contact predictionprotein complex structure predictionprotein language modelsgeometric deep learningprotein–protein docking |
| spellingShingle | Yunda Si Chengfei Yan Protein language model-embedded geometric graphs power inter-protein contact prediction eLife protein–protein interaction inter-protein contact prediction protein complex structure prediction protein language models geometric deep learning protein–protein docking |
| title | Protein language model-embedded geometric graphs power inter-protein contact prediction |
| title_full | Protein language model-embedded geometric graphs power inter-protein contact prediction |
| title_fullStr | Protein language model-embedded geometric graphs power inter-protein contact prediction |
| title_full_unstemmed | Protein language model-embedded geometric graphs power inter-protein contact prediction |
| title_short | Protein language model-embedded geometric graphs power inter-protein contact prediction |
| title_sort | protein language model embedded geometric graphs power inter protein contact prediction |
| topic | protein–protein interaction inter-protein contact prediction protein complex structure prediction protein language models geometric deep learning protein–protein docking |
| url | https://elifesciences.org/articles/92184 |
| work_keys_str_mv | AT yundasi proteinlanguagemodelembeddedgeometricgraphspowerinterproteincontactprediction AT chengfeiyan proteinlanguagemodelembeddedgeometricgraphspowerinterproteincontactprediction |