KATMAP infers splicing factor activity and regulatory targets from knockdown data.

KATMAP infers splicing factor position-specific regulatory activity and predicted cis-elements from SF perturbation RNA-seq plus binding motifs, enabling exon-level target predictions.

• Why it matters: RNA-seq reveals many splicing changes, but linking them to specific splicing factor mechanisms and direct targets remains difficult, especially when effects are indirect.
• What they did: KATMAP is an interpretable regression model that learns regulatory descriptions from SF knockdown RNA-seq data and an SF binding motif, using pretrained models on ENCODE SF knockdown datasets.
• The result: The learned models can predict SF regulation and cis-elements at individual exons, help distinguish direct versus indirect targets, and infer relevant SFs from clinical RNA-seq data.

AI summary of the abstract · may contain errors · not medical advice · read the paper.