Development and validation of an AI/ML platform for the discovery of splice-switching oligonucleotide targets

This study demonstrates the value that artificial intelligence/machine learning (AI/ML) provides for the identification of novel and verifiable splice-switching oligonucleotide (SSO) targets in-silico . SSOs are antisense compounds that act directly on pre-mRNA to modulate alternative splicing (AS). To leverage the potential of AS research for therapeutic development, we created SpliceLearn™, an AI/ML algorithm for the identification of modulatory SSO binding sites on pre-mRNA. SpliceLearn also predicts the identity of specific splicing factors whose binding to pre-mRNA is blocked by SSOs, adding considerable transparency to AI/ML-driven drug discovery and informing biological insights useful in further validation steps. Here we predicted NEDD4L exon 13 ( NEDD4Le13 ) as a novel target in triple negative breast cancer (TNBC) and computationally designed an SSO to modulate NEDD4Le13 . Targeting NEDD4Le13 with this SSO decreased the proliferative and migratory behavior of TNBC cells via downregulation of the TGFβ pathway. Overall, this study illustrates the ability of AI/ML to extract actionable insights from RNA-seq data. SpliceLearn is part of the SpliceCore® platform, an AI/ML predictive ensemble for AS-based drug target discovery.