Pablo Baeza-Centurión

The paper uses deep learning to predict splicing regulatory effects and identify effective antisense oligonucleotides by analyzing systematic deletion scans of human exons.

The paper shows that mutations rarely affect the inclusion of highly-included exons, suggesting disease-causing variants are more likely to affect alternatively spliced exons with intermediate inclusion levels.

DiMSum is a pipeline for analyzing deep mutational scanning data that provides variant fitness estimates and error modeling to identify experimental issues.

The paper shows that changes in gene expression alter the strength and direction of genetic interactions between mutations in many genes.