Multiple probabilistic models extract features from protein sequence data and resolve functional diversity of very different protein families

Abstract

Sequence functional classification has become a fundamental bottleneck to the understanding of the myriad of protein sequences accumulating in our databases due to the recent progress in genomics and metagenomics. The large diversity of homologous sequences hides, in many cases, a variety of functional activities that cannot be anticipated. Their identification appears critical for a fundamental understanding of living organisms and for biotechnological applications.ProfileView is a novel computational method designed to functionally classify sets of homologous sequences. Its architecture strongly relies on the structure of biological data, and answers to the challenge of automatically partitioning datasets of protein sequences in pertinent subfamilies based on meaningful conservation patterns. It constructs a library of probabilistic models accurately representing the functional variability of protein families, and extracts biologically interpretable information from the classification process. It applies to protein families that are not necessarily large, nor conserved, whose homologs might be very divergent and for which functions should be discovered or characterised more precisely.As a proof of concept, we apply ProfileView to the Cryptochrome/Photolyase family (CPF) and to the WW domain family, two widespread classes of proteins showing a large variety of functions and high sequence divergence. Decades of experimental studies on these families, functionally characterizing sequences and highlighting constitutive motifs, allow us to validate the two functional organisations obtained with the ProfileView approach. In addition, the method allows to identify a distinct functional group for the CPF, likely corresponding to novel photoreceptors. Thus, ProfileView appears as a powerful tool to classify protein sequences by function, screen sequences towards the design of accurate functional testing experiments and, possibly, discover new functions of natural sequences.Software and data availability: http://www.lcqb.upmc.fr/profileview/ A restrained access is momentarily set. Once the article is accepted, all information will be freely accessible.

MyCLADE: a multi-source domain annotation server for sequence functional exploration

“…Second, protein function for a combination of domains is not easily deducible based on the functions of single domains. Yet, the use of known domain combinations and the access to GO-terms for single domains can help the user for a fast exploration of potential protein functions ( 13 ) with MyCLADE.…”