Conservation signatures and sequence motifs support LaBranchoR predictions. (A) Average PhastCons and PhyloP conservation signatures centered on LaBranchoR predictions matching a Mercer et al. (2015). high confidence site (Match), matching a Mercer et al. (2015) low confidence site (Match Low Conf), 1–4 nt from a Mercer et al. (2015) site (<4 off, predicted), and more than 4 nt from a Mercer et al. (2015) site (>4 off, predicted), as well as centered on Mercer et al. (2015) sites 1–4 nt from a predicted site (<4 off, Mercer) and more than 4 nt from a LaBranchoR prediction (>4 off, Mercer). (B) Equivalent figure for Taggart et al. (2017) branchpoints. (C) The unusual 2′-OH linkage present at branchpoints can result in reverse transcriptase skipping over nucleotides near the branchpoint, causing the raw experimental data to differ from the true branchpoint. LaBranchoR predictions shifted 1 and 2 nt toward the 3′SS of Mercer et al. (2015) sites have the expected sequence motifs and PhyloP conservation signature. (D) Taggart et al. (2017) branchpoints often differ from LaBranchoR predictions by small shifts toward the 3′SS. LaBranchoR predictions shifted 1 and 2 nt away from the 3′SS of Taggart et al. (2017) sites have the expected sequence motifs and PhyloP conservation signature.
