Protein Contact Map Denoising Using Generative Adversarial Networks
Maddhuri Venkata Subramaniya, S. R.; Terashi, G.; Jain, A.; Kagaya, Y.; Kihara, D.
Show abstract
Protein residue-residue contact prediction from protein sequence information has undergone substantial improvement in the past few years, which has made it a critical driving force for building correct protein tertiary structure models. Improving accuracy of contact predictions has, therefore, become the forefront of protein structure prediction. Here, we show a novel contact map denoising method, ContactGAN, which uses Generative Adversarial Networks (GAN) to refine predicted protein contact maps. ContactGAN was able to make a consistent and significant improvement over predictions made by recent contact prediction methods when tested on two datasets including protein structure modeling targets in CASP13. ContactGAN will be a valuable addition in the structure prediction pipeline to achieve an extra gain in contact prediction accuracy.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- QDeep: distance-based protein model quality estimation by residue-level ensemble error classifications using stacked deep residual neural networks 96%
- Unsupervised protein embeddings outperform hand-crafted sequence and structure features at predicting molecular function 96%
- Scoring Protein Sequence Alignments Using Deep Learning 96%
Similar papers in this journal
- High-accuracy protein complex structure modeling based on sequence-derived structure complementarity 96%
- A Fast Approach for Structural and Evolutionary Analysis Based on Energetic Profile Protein Comparison 96%
- Accurate prediction of protein assembly structure by combining AlphaFold and symmetrical docking 96%
Similar papers in this journal
Similar papers in this journal
- GraphCPLMQA: Assessing protein model quality based on deep graph coupled networks using protein language model 96%
- Improved inter-protein contact prediction using dimensional hybrid residual networks and protein language models 95%
- Predicting the structures of cyclic peptides containing unnatural amino acids by HighFold2 95%
Similar papers in this journal
- Deep Template-based Protein Structure Prediction 97%
- Deducing high-accuracy protein contact-maps from a triplet of coevolutionary matrices through deep residual convolutional networks 96%
- Hybridized distance- and contact-based hierarchical structure modeling for folding soluble and membrane proteins 96%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.