Application of hierarchical function prediction in Solanum Lycopersicum

SPETALE, FLAVIO; ARCE, DÉBORA P; KRSTICEVIC, FLAVIA J; MURILLO, JAVIER; TAPIA, ELIZABETH; BULACIO, PILAR

IFMBE PROCEEDINGS

Springer. Huddinge : International Federation for Medical & Biological Engineering

Año: 2015 vol. 49

1680-0737

Predicting functional gene annotations through machine learning techniques may focus on the experimental validation reducing their cost. The hierarchical prediction method bases on true path rule (TPR) provides function results consistent and traceable with the Gene Ontology Molecular Function definition. In this work, a design of a hierarchical predictor model based on TPR for plants is presented. The training stage is done with Arabidopsis thaliana data characterized with sequence domains or physicochemical properties feeding an ensemble of binary classifiers, one classifier for each functional class. The proposed model is validated against a set of well-known control sequences and against a set of sequences of S. lycopersicum without any annotation. The discussed results are promising, the proposal can be extended and enriched with more organisms and with diverse sources of sequence characterizations.