Impact of phylogeny on structural contact inference from protein sequence data

Abstract

Local and global inference methods have been developed to infer structural contacts from multiple sequence alignments of homologous proteins. They rely on correlations in amino acid usage at contacting sites. Because homologous proteins share a common ancestry, their sequences also feature phylogenetic correlations, which can impair contact inference. We investigate this effect by generating controlled synthetic data from a minimal model where the importance of contacts and of phylogeny can be tuned. We demonstrate that global inference methods, specifically Potts models, are more resilient to phylogenetic correlations than local methods, based on covariance or mutual information. This holds whether or not phylogenetic corrections are used, and may explain the success of global methods. We analyse the roles of selection strength and of phylogenetic relatedness. We show that sites that mutate early in the phylogeny yield false positive contacts. We consider natural data and realistic synthetic data, and our findings generalize to these cases. Our results highlight the impact of phylogeny on contact prediction from protein sequences and illustrate the interplay between the rich structure of biological data and inference.

École polytechnique fédérale de Lausanne EPFL 83-66675-59-9

“…Our results regarding synthetic data suggest that this could be due to the robustness of these methods to phylogenetic correlations, which are bound to exist in natural sequence data. Here, we showed that phylogenetic correlations make the inference of functional sectors challenging, very much like they obscure the inference of structural contacts [4,[36][37][38][39][40]. It is important to note that phylogenetic correlations are nevertheless interesting and provide useful signal e.g.…”

“…in the case where all mutations are neutral. Phylogenetic correlations impair the inference of structural contacts from sequences [4,[36][37][38][39][40], which has motivated empirical corrections aiming at reducing their impact [4, 7-9, 41-43, 43-45]. Disentangling them from collectively correlated groups of amino acids such as functional sectors is bound to be a significant challenge too, perhaps even more.…”

Impact of phylogeny on the inference of functional sectors from protein sequence data

“…Their importance are controlled respectively by the parameters µ and κ: a smaller µ means that sequences are more closely related, yielding more phylogenetic correlations, while a larger κ means stronger selection, yielding more correlations arising from the sector. Note that this data generation process with phylogeny is close to the one we used previously [40,47], but that the Hamiltonian we use here (Eq 1) is specific to the sector model. B: To incorporate phylogenetic correlations, we start from one equilibrium sequence, which becomes the ancestor.…”

“…Our results regarding synthetic data suggest that this could be due to the robustness of these methods to phylogenetic correlations, which are bound to exist in natural sequence data. Here, we showed that phylogenetic correlations make the inference of functional sectors challenging, very much like they obscure the inference of structural contacts [4, 36–40]. It is important to note that phylogenetic correlations are nevertheless interesting and provide useful signal e.g.…”

“…in the case where all mutations are neutral. Phylogenetic correlations impair the inference of structural contacts from sequences [4, 36–40], which has motivated empirical corrections aiming at reducing their impact [4, 7–9, 41–43, 43–45]. Disentangling them from collectively correlated groups of amino acids such as functional sectors is bound to be a significant challenge too, perhaps even more.…”