The use of interactions of single nucleotide polymorphisms (SNPs), inherited genetic variants, to predict complex diseases is getting more attention during the past decade. However, clinical usage of these SNP-SNP interaction findings is low because few of identified SNP-SNP interactions can be validated and have known biological functions. This challenge can be improved by applying suitable statistical methods. Dr. Hui-Yi Lin, associate professor of biostatistics and lead author from Louisiana State University Health Sciences Center (LSUHSC) School of Public Health, New Orleans and her research team published a paper in Bioinformatics entitled “AA9int: SNP Interaction Pattern Search Using Non-Hierarchical Additive Model Set.”
[Photo: Dr. Hui-Yi Lin]
The AA9int approach is composed of nine interaction models by considering non-hierarchical model structure and the additive mode. The conventional approach for testing SNP-SNP interactions requires the hierarchical rule, which requires both main effects to be included in a model when testing their interaction, and this approach only tests one specific interaction pattern. In order to test more interaction patterns, the hierarchical rule needs to be loosened. For thoroughly searching SNP-SNP interactions associated with a complex disease, Dr. Lin previously developed a powerful statistical methods: the SNP Interaction Pattern Identifier (SIPI) [HY Lin et al. 2017 in Bioinformatics]. SIPI tests 45 patterns by considering non-hierarchical structure, inheritance mode (additive, dominant and recessive), and mode coding direction. For large-scale studies with thousands of SNPs, an effective and computation-efficient method is needed to serve as a screening method in the two-stage approach. The new proposed AA9int, a mini-version of SIPI, tests the nine additive patterns out of the full 45 patterns. In the simulation results, it is shown that non-hierarchical models play a more important role in SNP interaction detection than inheritance modes. AA9int has similar statistical power compared to SIPI and is superior to the conventional approach. In summary, it is recommended that AA9int is a powerful tool to be used either alone or as the screening stage of a two-stage approach (AA9int+SIPI) for detecting SNP-SNP interactions in large-scale genetic association studies.