Improved Modeling of RNA-binding Protein Motifs in an Interpretable Neural Model of RNA Splicing
Overview
Authors
Affiliations
Sequence-specific RNA-binding proteins (RBPs) play central roles in splicing decisions. Here, we describe a modular splicing architecture that leverages in vitro-derived RNA affinity models for 79 human RBPs and the annotated human genome to produce improved models of RBP binding and activity. Binding and activity are modeled by separate Motif and Aggregator components that can be mixed and matched, enforcing sparsity to improve interpretability. Training a new Adjusted Motif (AM) architecture on the splicing task not only yields better splicing predictions but also improves prediction of RBP-binding sites in vivo and of splicing activity, assessed using independent data.
Andrews R, Bass B bioRxiv. 2025; .
PMID: 39975386 PMC: 11838218. DOI: 10.1101/2025.01.24.634786.
Laine E, Freiberger M Curr Opin Struct Biol. 2025; 90:102979.
PMID: 39778413 PMC: 7617313. DOI: 10.1016/j.sbi.2024.102979.
Decoding biology with massively parallel reporter assays and machine learning.
La Fleur A, Shi Y, Seelig G Genes Dev. 2024; 38(17-20):843-865.
PMID: 39362779 PMC: 11535156. DOI: 10.1101/gad.351800.124.
Capitanchik C, Wilkins O, Wagner N, Gagneur J, Ule J Nat Rev Genet. 2024; 26(3):171-190.
PMID: 39358547 DOI: 10.1038/s41576-024-00774-2.
An interpretable model of pre-mRNA splicing for animal and plant genes.
McCue K, Burge C Sci Adv. 2024; 10(19):eadn1547.
PMID: 38718117 PMC: 11078188. DOI: 10.1126/sciadv.adn1547.