Folding membrane proteins by deep transfer learning

By Sheng Wang, Zhen Li, Yizhou Yu, Jinbo Xu. 2017

74 citations recorded by the source corpus.

Computational elucidation of membrane protein (MP) structures is challenging partially due to lack of sufficient solved structures for homology modeling. Here we describe a high-throughput deep transfer learning method that first predicts MP contacts by learning from non-membrane proteins (non-MPs) and then predicting three-dimensional structure models using the predicted contacts as distance restraints. Tested on 510 non-redundant MPs, our method has contact prediction accuracy at least 0.18 better than existing methods, predicts correct folds for 218 MPs (TMscore>0.6), and generates three-dimensional models with RMSD less than 4Å and 5Å for 57 and 108 MPs, respectively. A rigorous blind test in the continuous automated model evaluation (CAMEO) project shows that our method predicted high-resolution three-dimensional models for two recent test MPs of 210 residues with RMSD ∼2Å. We estimated that our method could predict correct folds for 1,345–1,871 reviewed human multi-pass MPs including a few hundred new folds, which shall facilitate the discovery of drugs targeting at membrane proteins.

Subfields: Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

A Review of Deep Transfer Learning and Recent Advancements (2022, methodology): “Even with implementation limitations (freezing part of a layer), our evaluation demonstrates the high potential of the EXPANSE approach for a successful transfer in deep learning that can be applied to different DL models and tasks.”
De novo Protein Structure Prediction by Coupling Contact With Distance Profile (2020, methodology): “In RaptorX [18], [26], [27], [28], predicted contacts are utilized as distance constraints to predict tertiary structural models.”
Deep Learning in Protein Structural Modeling and Design (2020, methodology): “SPIN209 reached a 30.7% sequence recovery rate and Wang et al.194 and O’Connell et al.’s25 SPIN2 further improved it to 34%.”
ComplexContact: a web server for inter-protein contact prediction using deep learning (2018, methodology): “According to our experience on membrane protein contact prediction (39), using a mix of membrane and soluble proteins to train a DL model works better than using soluble proteins or membrane proteins alone.”
Analysis of deep learning methods for blind protein contact prediction in CASP12 (2017, methodology): “We are not aware of any CASP12 hard targets being membrane proteins, but our deep learning method produced decent models for 4 membrane proteins in the blind CAMEO test.(26,27) This finding implies that the sequence-structure relationship learned by our model from soluble proteins can be transferred”