In the context of health analytics, the challenge was to discover which Single-nucleotide polymorphisms (SNPs) or combinations of them are involved in the development of Gestational Diabetes Mellitus (GDM), together with the influence of a number of patient basic clinical data. Additionally, an algorithm to quickly determine the odds of suffering GDM prior to pregnancy based on these genetic data is needed.
Identification of individual and groups of SNPs relevant for early GDM detection, and a model to produce a fine-tuned score that predicts the a prior odds of suffering such condition. A detailed analysis of the SNP combinations that either protect or expose to a high risk of GDM was also provided.
Some SNPs stood out on various metrics, and our model identified and quantified several risk factors that can be spotted a priori.