KEDROS: A Comparative Genomics and High-Resolution Phylogenetic Profiling Workspace
Phylogenetic profiling infers gene function from the presence and absence of genes across genomes, but it loses resolution in eukaryotes, where many protein families contain several functionally distinct paralogs. Existing resources address parts of this problem separately: web services build trees for one family at a time, orthology databases provide precomputed orthologous groups for predefined…
Phylogenetic profiling is a method used to infer gene function by examining the presence and absence of genes across various genomes. However, it becomes less precise in eukaryotes, as many protein families have multiple functionally distinct paralogs. Currently, there are separate resources that tackle different aspects of this problem: web services construct trees for individual protein families, orthology databases offer precomputed orthologous groups for certain species sets, and tree-profiling tools necessitate users to provide their own trees and groupings.
Researchers have now developed KEDROS, a web-based platform that addresses the entire workflow from start to finish.
KEDROS allows users to build gene trees for any protein family across UniProt reference proteomes. It then splits each tree into subfamilies based on inferred gene duplications. Users can choose the organisms, protein families, and taxonomic depth at which trees are split, enabling KEDROS to tackle a wide range of comparative questions. These questions can involve various contrasts, such as a pathogen and its host, a pest and non-target species, or lineages with and without a specific trait.
The platform provides two levels of analysis. First, it ranks gene families and subfamilies present in one set of organisms but absent from another, highlighting lineage-specific genes and potential targets. Within shared subfamilies, KEDROS identifies amino acid positions that differ consistently between organisms, which could underlie functional divergence. It displays these differences on AlphaFold structures, aiding in the visualization of protein changes.
All of these analyses can be performed directly in a web browser, requiring no installation or programming knowledge. While similar functionalities exist elsewhere, no other resource combines them into a single, end-to-end solution at the scale of whole proteomes, allowing users to define their own organism sets. To demonstrate the power of KEDROS, the researchers present two case studies.
In the first case, they examine the trypanothione reductase family in trypanosomatids, uncovering parasite-specific subfamilies and substrate-binding substitutions that differentiate the parasite enzyme from human glutathione reductase. In the second case, they analyze the bacterial phylum Bacillota, using only phenotype labels to recover the sporulation machinery across 1,218 proteomes. KEDROS also ranks an uncharacterized family among the top markers of the sporulation trait.
In summary, KEDROS offers a user-friendly, code-free approach to transitioning from a comparative biological question to testable hypotheses about genes and residues. It is freely accessible at https://kedros.cgenomicslab.org.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.