Identification of a novel UPR related pathway through phylogenetic profiling-based network analysis

Most human genes have limited characterization. It is reasonable to assume that some of these genes participate in currently unidentified pathways. Phylogenetic profiling is a comparative genomics approach based on the assumption that genes working together undergo similar evolutionary pressures. Therefore, if we look at the conservation of a genes across multiple species, genes with similar conservation profiles are likely to functionally interact. We aimed to use this unbiased method to connect genes, assign new functions to uncharacterized genes, and identify novel pathways. We analyzed the 20,000 human genes across 1905 species, divided into 12 clades (e.g., Mammalia, Fungi). Following normalizations, we integrated the information from the different clades to build the co-evolution gene network. We then applied the DREAM challenge TUSK algorithm to identify highly connected regions within the network. These regions, also referred to as communities or clusters, include ~26% known pathways, and the rest are clusters of genes with unknown functions. We chose one of these clusters to validate our computational approach. The cluster was predicted to reside in the ER, so we proceeded to show the role of the genes in the untranslated protein response (UPR) in C. elegans. We were able to show these genes indeed effect UPR, specifically through the ATF6 pathway. We also showed that down regulation of these genes caused constant activation of UPR. We present a database containing all the identified clusters along with data integration from multiple databases to direct the search for the role of the human uncharacterized genes.