HPC MSU

Publication Abstract

Top-down transversing the Gene Ontology to extract biological knowledge for biomedical models

Buza, T., Arick, M., II, Gresham, C. R., McCarthy, F.M., & Nanduri, B. (2013). Top-down transversing the Gene Ontology to extract biological knowledge for biomedical models. MCBIOS X: The 10th Anniversary: Discovery in A Sea of Data. University of Missouri, Columbia, MO: MidSouth Computational Biology and Bioinformatics Society (MCBIOS). 10(1), 120.

Abstract

Modern molecular biology research is moving from high-input, low-output to low-input, high-throughput ‘omics data generation. Simultaneously biologists face the challenge of converting this data into high-level biological knowledge. Due to limited resources, few model genomes are comprehensively annotated to serve as references, especially for studying human diseases. Limited resources hinder development of new animal models for human diseases. Developing a highly curated, cross-referenced resource that links key features of human disease causing genes to non-model animal genes and provision of computational methods for mining this resource provides opportunities to understand human-non-model animal disease relationships. Gene Ontology (GO) terms are neutral key features that provide functional information about relationships of disease genes. GO terms are organized in Directed Acyclic Graphs (DAG), in which a term can be a child of one or more parents in a hierarchical relationship. The terms provide more general to most specific biological knowledge when transversing hierarchically from root to leaf. A gene product may be annotated to different levels in the DAG, making it difficult for users to pinpoint the most specific function of a gene. Here, we develop a tool that integrates the most specific experimentally verified GO terms for human disease causing genes, their curated pathways, phenotypes and genetic disorders with orthologous genes in cow, a non-model animal. This initial work provides a foundation for developing a highly curated, cross-referenced resource for studying human diseases in non-model animals.