Bayesian network analysis of SNPs rs11153071 and rs2107597 and clinical risk factors in intracranial aneurysms

Objective: Intracranial aneurysms (IAs) are abnormal dilations of blood vessels in the brain that, especially if ruptured, can lead to death or disability. Around 3% of people are affected, with genetic and environmental factors contributing to the aetiology. The objective of this analysis is to develop a model quantifying the interrelationship between genetic and non-genetic factors that define IA, utilizing two single-nucleotide polymorphisms (SNPs) which presumably affect phenotypic switching of vascular smooth muscle cells. Methods: The ISGC-IA Working Group’s data, collected over the past two decades on 2,105 patients with a ruptured IA of any size, was analysed, including patients and IA characteristics. A Directed Acyclic Graph (DAG) was constructed to illustrate relationships between the evaluated variables applying an additive Bayesian network (BN) using the R-package abn. Further, multiple regression analyses were performed. Findings: Preliminary results from the BN show no influence of either SNP on any of the observed variables, see Figure 1. Regression analyses showed moderate evidence for an effect of rs11153071 on positive family history (OR = 1.30, 95% CI: 1.03 to 1.63, p−val = 0.024) and on IA location (OR: 1.24, 95% CI: 1.05 to 1.46, p−val = 0.011). No association was found for rs2107597 genotypes. Consequently, the BN model exhibited no association between patient or IA characteristics and SNP genotypes, but regression analysis identified evidence of an effect from one SNP. Modelling of the multicentric nature of the data and inclusion of unobserved confounders can facilitate enhanced comprehension of the interrelatedness of genetic and non-genetic risk factors of the disease.