A Study on Frequent Itemset Mining for Identifying Associated Multiple SNPs

Genome-wide association studies (GWAS) have gained a lot of interest in public health research to investigate the correlations of genetic variants and traits. Mostly, GWAS use standard statistical tests for each genetic variant to capture main genetic effects. Machine learning and data mining approaches are also promising enough to complement single and multiple genetic variants in understanding the general association of complex human disease. This paper explores a data mining approach to discover patterns of multiple genetic variants associated with a disease. Frequent itemset mining method was applied and the intersection algorithm in a row enumeration strategy was chosen to discover itemsets from genetic variants, which is known as Single Nucleotide Polymorphism (SNP). We chose the intersection algorithm because it is more suitable to mine high dimensional and sparse dataset. The found itemsets could be used by scientists to study associated with multiple genes in multifactorial disease.