A Subgroup Discovery Approach for Relating Chemical Structure and Phenotype Data in Chemical Genomics
Lan Umek, Petra Kaferle, Mojca Mattiazzi Ušaj, Aleš Erjavec, Črtomir Gorup, Tomaž Curk, Uroš Petrovič, Blaž Zupan · 2010
We report on development of an algorithm that can infer relations between the chemical structure and biochemical pathways from mutant-based growth tness characterizations of small molecules. Identication of such relations is very important in drug discovery and development from the perspective of argument-based selection of candidate molecules in target-specic screenings, and early exclusion of substances with highly probable undesired side-eects. The algorithm uses a combination of unsupervised and supervised machine learning techniques, and besides experimental tness data uses knowledge on gene subgroups (pathways), structural descriptions of chemicals, and MeSH term-based chemical and pharmacological annotations. We demonstrate the utility of the proposed approach