Summary
Multifactorial Study for the Identification of New Genetic and Non-Genetic Factors Using Exome Sequencing and Artificial Intelligence Techniques
Description
Type 2 diabetes (T2D) causes a significant increase in morbidity and premature mortality. The genetic component of T2D is over 50%, but it remains largely unidentified. Our main objective is to identify the genetic variations, genes, environmental factors, and their interactions involved in T2D.
Methodology:
- Identification of all genetic variants present in the exome of participants in the Di@bet.es study who were diagnosed with T2D at the start of the study, elderly non-diabetics, and all individuals with follow-up after 7.5 years (approximately 3,400 participants).
- Analysis of all the variations found in relation to T2D and related parameters using conventional statistics (BMI, insulin and glucose levels, metabolic syndrome, insulin resistance, response to glucose load, lipid profile, and evolution after 9 years in all these parameters).
- Application of artificial intelligence (machine learning) techniques to: a) Find genetic variants involved in T2D and related parameters; b) Identify environmental factors associated with T2D; c) Identify interactions between multiple influential variants in high-dimensional spaces (tens of thousands of genetic variants) and between them and other non-genetic factors; d) Create predictive models for T2D based on these influential factors; e) Test models and approaches with the diagnosis at the start of the study and quantify their ability to predict who developed T2D during follow-up, and in 800 exomes (cases and controls) sequenced by our group. Additionally, model and test around 13,000 exomes obtained from public databases (from the TD2-GENES and GoT2D consortia).