Supervised Classification of Genetic Sequences for Population Analysis

We used a Support Vector Machine (SVM) algorithm for analysis of biological populations through supervised classification of sequence data. We describe an open-source software which implements the method, and apply the method and the program to analyze environmentally challenged populations of the estuarine fish Fundulus heteroclitus. Specifically, we investigate whether the genetic composition (DNA sequence) of a particular detoxification locus predicts population assignment of fish to chemically contaminated versus clean estuaries. The analysis method uses an SVM algorithm to assign individual fish, characterized by their allelic composition, into a toxic-resistant or ∗The corresponding author. Mailing address: San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132. E-mail: [email protected]. non-resistant group. We employed classification error in assignment as a measure of population similarity. The results validate the proposed method by providing supporting evidence for the previously suggested role of AHR1 (aryl hydrocarbon receptor) locus in the toxic response pathway of Fundulus heteroclitus.