Penalised neural network with weight correlation descent for predicting polygenic risk score
Kim, S. b.; Kang, J. H.; Cheon, M.; Kim, D. J.; Lee, B.-C.
Show abstract
In this study, we developed a deep neural network (DNN) model with weight correlation descent (WCD) regularization to improve polygenic risk score predictions for complex diseases, specifically gender-specific cancers, using the UK Biobank dataset. Our DNN model with WCD outperformed both conventional PRS models and DNN models without WCD, demonstrating the importance of regularization techniques in enhancing model performance and capturing non-linear effects and interactions in genomic data. These findings contribute to a better understanding of genetic architecture, facilitating personalized interventions based on individual genetic profiles and ultimately benefiting patient care and health outcomes.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.