Figure 1. The human splicing code.

(a) For a given cell type, the computational model extracts the regulatory code from a test DNA sequence and predicts the percent of transcripts with the exon spliced in, Ψ. (b) Predictions were made for 10,689 test exons profiled in 16 tissues, exons and tissues were binned according to their RNA-seq assessed values of Ψ, and for each bin (column) the distribution of code-predicted Ψ is plotted (n=56,104).