Prediction of gene expression from transcription factors affinities: an application of Bayesian non-linear modelling

Abstract

The prediction of gene expressions from DNA sequences is a relevant problem in biology. While most of the existing methods dedicated to this task use genotypes as predictors, here we propose a method based on transcription factor affinities, which have a clearer biological interpretation. This novelty, however, introduces new challenges for modelling, which we address leveraging on Bayesian non-linear modelling techniques.

Silvia Montagna

Assistant Professor in Statistics