Extended Data Fig. 7: Modelling SAV effects on the canonical transcript.
From: Resurrecting the alternative splicing landscape of archaic hominins using machine learning
We used the SpliceAI output to construct a novel transcript per SAV by modifying the canonical transcript for that gene. We considered only one effect per SAV (for example, either an acceptor gain, acceptor loss, donor gain or donor loss) based on the effect with the largest ∆. Therefore, we did not model multiple effects for a single SAV (for example, an acceptor gain and acceptor loss). Here, we illustrate all the possible consequences of a SAV for each of the four classes. We indicate the variant position with a red ‘X’ and the position of the effect with a red vertical line (sequence deletion) or box (sequence addition). Each scenario includes a two exon gene (boxes) with a single intron (horizontal line).
