genetic classification of azari and north ecotype buffalo population using svm method

the purpose of this research was to classify buffaloes from different areas of the two azari (west and east azarbayjan and ardabil provinces) and north (guilan province) ecotypes using support vector machine method. a total of 258 buffalo were sampled and genotyped using the axiom buffalo 90k genotyping array at the parco technologic padano lab in italy. two metric methods of cross validation and the area under the receiver operating characteristic (auc) were used to determine the predictive performance of support vector machine (svm) to classify individuals. the results of cross validation and methods for classifying different regions of the two ecotypes (4 provinces) were 92% and 96%, respectively that showed despite the difficulty of identifying individuals from provinces close to each other, support vector machine (svm) method shows higher accuracy in assigning animals to their herds. result of two ecotypes showed accuracy about 96% and 98% which represents the better ability to separate the two ecotypes. machine learning method provides predictions for classification of each individual which can be efficient in quality control and genetic studies.