The Regional Program STIC-AmSud is an initiative of the French cooperation and their counterparts from Argentina, Brazil, Chile, Paraguay, Peru and Uruguay, aimed to promote and strengthen the collaboration and to create networks of research and development in the field of Information and Communication Science and Technology (ICT), through joint projects.
This project is oriented to the design of a knowledge-based model for the prediction and characterization of associations between Single Nucleotide Polymorphisms (SNPs) and phenotypic traits. Strategically, our proposal is based on the characterization of SNPs interactions by supervised machine learning techniques enriched and guided by expert knowledge, i.e., putative positions of mutations, candidate gene sets, inheritance issues, and epigenetic evidence. The broad characterization of SNPs interaction at gene and intergene levels is necessary to infer complex traits. The inclusion of expert knowledge is fundamental to reduce the computational complexity of such broad characterizations. Under this baseline, a prototype model for predicting eye color defective traits arising from SNPs in eye pigment genes of the fruit fly Drosophila melanogaster is considered. This project aims to contribute to a long-term objective of understanding complex traits, i.e., multigenic disorders. For this purpose, the development of machine learning methods for modeling expert knowledge and information sources are considered. In particular, the problem of predicting SNP-phenotypic trait associations involves noisy and multi-dimensionally correlated high dimensional datasets often lacking of the appropriate number of samples. The effective treatment of these information sources requires the integration of complex expert knowledge, a challenging issue to which the results of this project could bring an adequate solution.