High order interaction analysis of SNPs in PEDF (rs12150053, rs12948385) and EPO (rs1617640) genes with clinical determinants of type 2 diabetic retinopathy patients from south India

- Significant difference was not seen in the distribution of SNPs rs12150053, rs12948385 and rs1617640 between DR and No DR group.
- The CC genotype of rs12150053 in PEDF gene associated with DR after adjusting for insulin user status
- Significant interaction for glycemic control (denoted by HbA1c levels) and blood pressure parameters with rs12150053 in both single and additive models.

Gene environment interaction in complex diseases like type 2 diabetes retinopathy (DR) may provide valuable insight into the complexity of disease. Current study aims to explore the gene-gene and gene-environment interaction in the pathology of type 2 DR. SNPs in PEDF (rs12150053, rs12948385) and EPO (rs1617640) genes are studied for the potential interactions with various clinical risk factors in type 2 diabetes patients of south Indian origin with (N = 201) and without (N = 168) retinopathy. The interaction study is performed with the statistical tools namely generalized multifactor dimensionality reduction (GMDR) and classification and regression tree (CART) methods. GMDR analysis showed a probable interaction for insulin user status with the 3 SNP that remained significant after adjusting for various clinical factors [Testing balance accuracy of 62.41 and 61.49 respectively]. PEDF polymorphisms was also seen interacting with HbA1c levels (p < 0.05; TBA > 59). SNP rs12150053 determined the subsequent split among insulin users and poor control of HbA1c by CART. The CC genotype of rs12150053 showed an OR = 4.9 after adjusting for insulin user status. We did not find any direct disease association for the SNPs with DR in the study population. The study showed insulin user status and glycemic index as the probable interacting factors with DR, potentially modified by rs12150053. However, the direct role of these SNPs in regulating these interaction demands functional validation and replication for statistical significance.