A simple and efficient way to calculate gene set signature scores for use in clinical classification models

Polygenic score is a term borrowed from the field of genome-wide association studies (GWAS). In GWAS, most of the susceptibility SNPs are associated with small risk estimates, hence they have little predictive values. However, when multiple variants are combined into a single risk score, they can serve as a good estimate of an individual`s genetic liability to a trait or disease. A similar approach can be used in other omics fields to calculate a single score from a set of genomic features that can serve as a prognostic, diagnostic, or treatment response marker with greater predictive power than each genomic feature separately. Here I’ll present a few strategies to calculate gene set signature scores based on either transcriptomics or epigenomics data and will demonstrate their use in cancer and psychiatry fields respectively. Specifically, using the polygenic approach, we developed a gene set signature score associated with reactive stroma in HER2-positive early breast cancer that may predict resistance to adjuvant trastuzumab therapy. In another study, we established a putative polygenic DNA methylation score in cord blood from infants that predicts maternal depression. In summary, the omics-based polygenic score approach is an efficient and simple method to integrate genome-wide information in a manageable number of dimensions, and it may facilitate experimental studies that examine the multi-omics basis of human complex traits.