EPFL BioE Talks SERIES "Accounting for Phylogeny in Protein Coevolution Improves Structural Contact Prediction and Reveals Residue Clusters With Different Evolutionary Histories"


Event details

Date 23.11.2020 16:0016:30  
Speaker Prof. Kerwyn KC Huang, Stanford University, Stanford, CA (USA)
Location Online
Category Conferences - Seminars
(note that this talk is number one of a double-feature seminar - see details of the second talk here)

It has long been appreciated that comparisons among homologous sequences of a protein of interest can reveal residues key to its function and/or structure, and just as evolutionarily conserved individual residues are generally crucial to proper protein function, the statistical covariation (coevolution) between pairs of residues has been shown to carry information such as functional subdomains and structural contacts. However, coevolution measurements are often obscured by genetic drift, which defines a protein’s phylogenetic history and gives rise to concurrent variation of positions across the protein sequence. Here, we describe a robust method for explicitly separating the phylogenetic dimension of coevolution signal, and demonstrate that coevolution can occur on multiple phylogenetic timescales within a single protein. Our method, called nested coevolution (NC), can be applied as a correction to any coevolution metric and predicts that poorly conserved residues can nonetheless have important roles in protein function. NC improves structural contact prediction over gold-standard methods, particularly in subsampled alignments with fewer sequences, and lowers the noise in detection of sectors of coevolving residues. We show that the NC sectors are spatially compact, phylogenetically distinct from the rest of the protein in a conservation-independent manner, and predict mutations that disrupt protein activity. Our conceptualization of the phylogenetic separation of coevolution represents a departure from previous attempts to reduce phylogenetic noise, and enables broad application of protein coevolution measurements, particularly to naturally sparse eukaryotic proteins, with the potential to elucidate the relationships between protein evolution and genetic diseases.

KC Huang was an undergraduate Physics and Mathematics major in Page House at Caltech, and spent a year as a Churchill Scholar at Cambridge University working with Dr. Guna Rajagopal on Quantum Monte Carlo simulations of water cluster formation. He received his PhD from MIT working with Prof. John Joannopoulos on electromagnetic flux localization in polaritonic photonic crystals and the control of melting at semiconductor surfaces using nanoscale coatings. During a short summer internship at NEC Research Labs, he became interested in self-organization in biological systems, and moved on to a postdoc with Prof. Ned Wingreen in the Department of Molecular Biology at Princeton working on the relationships among cell shape detection, determination, and maintenance in bacteria. His lab is currently situated in the departments of Bioengineering and Microbiology & Immunology at Stanford, and his current interests include cell division, membrane organization, cell wall biogenesis, and collective motility of bacterial communities. His lab is a force to be reckoned with in IM basketball.

Zoom link (with registration) for attending remotely: https://go.epfl.ch/EPFLBioETalks

IMPORTANT NOTICE: due to restrictions resulting from the ongoing Covid-19 situation, this seminar can be followed via Zoom web-streaming only, following prior one-time registration through the link above.

Practical information

  • Informed public
  • Registration required