Machine learning based CRISPR gRNA design for therapeutic exon skipping
Fig 1
Genome-integrated reporter system, and machine learning pipeline for predicting exon skipping levels from SpCas9 gRNAs that target splice acceptors.
(A) A simplified illustration of the genome-integrating high-throughput CRISPR-Cas9 splicing reporter system. (B) MetaSplice combines predictions from SpliceAI, and MMSplice’s intronic, acceptor site, and exonic modules using a linear model tuned on experimental data from (A). (C) The full machine learning pipeline, SkipGuide, predicts ΨG for a given SpCas9 gRNA, by chaining inDelphi and MetaSplice. Probabilistic interpretations are shown in red.