Large-Scale Orthology Predictions for Inferring Gene Functions Across Multiple Species
ARMY MEDICAL RESEARCH AND MATERIEL COMMAND FORT DETRICK MD
Pagination or Media Count:
An effective approach to infer the functions of genes is to use the concept of gene orthology. Because orthologous genes are likely to share similar functions the functions of genes in an unstudied species can be inferred through the functions of their orthologs in a studied model species. To infer gene functions for a multitude of species, we developed a high-throughput orthology prediction method, termed PhyloTrace. PhyloTrace is both highly accurate and computationally efficient for large-scale applications, having the ability to infer orthologous genes across thousands of species. This is accomplished through three major steps 1 allagainst- all gene comparisons for every pair of genes, 2 pair-wise orthology predictions for every two genomes and 3 the generation of orthologous clusters that contain orthologous genes across multiple genomes. We employed the previously developed Pipeman parallelization program to break down a set of millions of input sequences into small chunks and then processed them in parallel. We successfully predicted orthologs for over 900 bacterial genomes, achieving a falsepositive prediction rate of 2.0, which was a significant improvement compared with the widely used bidirectional best-hit method, which yielded a falsepositive rate of 5.5.
- Genetic Engineering and Molecular Biology
- Medicine and Medical Research