Fig. 1

Quantifying changes in alternative polyadenylation with LABRAT. (A) LABRAT computational pipeline. (B) Explanation of ψ as a metric of polyadenylation site choice. Genes that exclusively use upstream or gene-proximal sites have ψ values of 0 while those that exclusively use downstream or gene-distal sites have ψ values of 1. The two transcript structures associated with alternative polyadenylation, tandem UTRs and alternative last exons, are diagrammed. (C) Comparison of ψ values in mouse brain and liver RNA for genes whose ψ value was significantly different between these tissues. (D) RNA coverage profiles of a gene with differential polyadenylation site usage in mouse brain and liver tissues. Dots represent ψ values calculated in each of 8 replicates. (E) RNA coverage profile of a gene with differential polyadenylation site usage in control PBMCs and those treated with poly dI:dC. RNA from these cells was profiled using 3′ end sequencing. Dots represent ψ values calculated in each of 3 replicates. (F) Comparison of ψ values between RNA samples profiled using standard RNAseq libraries (purple) and 3′ end sequencing libraries (orange). RNAseq samples were quantified by supplying ‘RNAseq’ as the ‘librarytype’ parameter for LABRAT while 3′ end sequencing libraries were quantified by supplying ‘3pseq’ as the ‘librarytype’ parameter. (G) Benchmarking of LABRAT performance against other widely used software package for quantification of alternative polyadenylation from RNAseq data