Studying the genetic factors behind complex diseases is challenging because they are influenced by many genes and their interactions with each other and the environment. While genome-wide association studies (GWAS) have identified many genetic variants linked to various traits, traditional methods often assume simple relationships and don’t account for interactions between genes (epistasis).
In this study, researchers used artificial neural networks (a type of machine learning) to predict complex traits using both simulated and real genetic data. They applied methods to identify which genetic regions were most important for predicting specific traits. Simulations showed that the method could accurately identify associated regions, but further analysis was needed to pinpoint specific genetic variants, similar to conventional GWAS.
Using this approach on data from people with schizophrenia in the Estonian Biobank, they identified multiple genetic regions potentially linked to the disorder. These regions were enriched in genes related to brain structure, aligning with the biological understanding of schizophrenia.
With further refinement, neural networks could become powerful tools for identifying genetic factors in complex diseases, complementing traditional GWAS methods and paving the way for more detailed genetic studies.
Access the article here.