Annotation pipeline for inferring gene functions integrating GO annotations and expression data

L. DI PERSIA, D.H. MILONE AND G. STEGMAYER

Mendoza

Congreso; X Congreso Argentino de Bioinformática y Biología Computacional; 2019

A2B2C

Background:Computational methodsfor the prediction of gene function refers to automatically findingassociations between a gene and a set of Gene Ontology (GO) terms.Since the hand-made curation process of novel annotations are verytime-consuming, computational tools that can reliably predict likelyannotations and boost the discovery of new gene annotations areurgently needed. Results:This work proposes anovel pipeline (see Figure) for inferring gene annotations based onthe automatic reconstruction of the semantic similarity betweengenes. The semantic similarity is a metric defined over a set ofterms, where the distance between them is based on the likeness oftheir meaning or semantic content. We benchmarked the proposalagainst state-of-the-art methods on three published data sets(Arabidopsis thaliana,Saccharomycescerevisiae andDictyosteliumdiscoideum).Independent experiments have shown that the proportion betweenannotated and unannotated genes does not influences the modelaccuracy.