The generative capacity of probabilistic protein sequence models
Open Access
- 2 November 2021
- journal article
- research article
- Published by Springer Nature in Nature Communications
- Vol. 12 (1) , 1-14
Abstract
Potts models and variational autoencoders (VAEs) have recently gained popularity as generative protein sequence models (GPSMs) to explore fitness landscapes and predict mutation effects. Despite encouraging results, current model evaluation metrics leave unclear whether GPSMs faithfully reproduce the complex multi-residue mutational patterns observed in natural sequences due to epistasis. Here, we develop a set of sequence statistics to assess the “generative capacity” of three current GPSMs: the pairwise Potts Hamiltonian, the VAE, and the site-independent model. We show that the Potts model’s generative capacity is largest, as the higher-order mutational statistics generated by the model agree with those observed for natural sequences, while the VAE’s lies between the Potts and site-independent models. Importantly, our work provides a new framework for evaluating and interpreting GPSM accuracy which emphasizes the role of higher-order covariation and epistasis, with broader implications for probabilistic sequence models in general.Keywords
All Related Versions
Funding Information
- National Science Foundation (1934848, 1625061)
- U.S. Department of Health & Human Services | National Institutes of Health (R35-GM132090)
- United States Department of Defense | United States Army | U.S. Army Research, Development and Engineering Command | Army Research Laboratory (W911NF-16-2-0189)
This publication has 66 references indexed in Scilit:
- Correlated Electrostatic Mutations Provide a Reservoir of Stability in HIV ProteasePLOS Computational Biology, 2012
- Genomics-aided structure predictionProceedings of the National Academy of Sciences, 2012
- HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignmentNature Methods, 2011
- Pairwise and higher-order correlations among drug-resistance mutations in HIV-1 subtype B proteaseBMC Bioinformatics, 2009
- Identification of direct residue contacts in protein–protein interaction by message passingProceedings of the National Academy of Sciences, 2009
- Evolutionary information for specifying a protein foldNature, 2005
- Network Information and Connected CorrelationsPhysical Review Letters, 2003
- Variable Selection via Nonconcave Penalized Likelihood and its Oracle PropertiesJournal of the American Statistical Association, 2001
- Correlated mutations in models of protein sequences: phylogenetic and structural effects1999
- DYNAMICS OF PROTEINS: ELEMENTS AND FUNCTIONAnnual Review of Biochemistry, 1983