The approach to patient clustering based on the microchip data confined to distinct loci using the combinations of variants

Fuchs' endothelial corneal dystrophy is a socially significant hereditary disease. More than a half of cases in the European population are caused by the increased number of trinucleotude repeats in the TCF4 gene. The study was aimed to develop and test the approach of dividing patients into groups based on the chip-based genotyping and genome-wide association study (GWAS) results. The analysis was conducted using FECD Genetics Multi-center Study and AREDs project datasets containing the data of 1721 clinical cases and 2408 control patients. When analyzing the GWAS results, the patients and the control group were divided into two groups by means of hierarchical clustering suggesting that patients with the increased number of repeats in the TCF4 gene are carriers of specific combinations of genomic variants (haplotypes). It was shown that individual variants cannot be used for the molecular genetic stratification of patients with the increased number of repeats in TCF4 due to inconsistent results obtained for the variants. Furthermore, the haplotype-based approach outperformed the SNPs in terms of odds ratio. The paper proposes a method that enables further search for the biologically relevant combinations of genomic variants.