Skip to main content

Table 5 Quality measures in detecting annotated alternative splicing events on the simulated datasets with 5M and 10M reads

From: ASGAL: aligning RNA-Seq data to a splicing graph to detect novel alternative splicing events

   5M 10M
Tool Measure ES A3 A5 IR ES A3 A5 IR
ASGAL Prec 0.999 0.850 0.702 0.657 0.999 0.846 0.703 0.642
  Rec 0.924 0.788 0.774 0.719 0.966 0.814 0.799 0.722
  FM 0.960 0.818 0.736 0.687 0.982 0.830 0.748 0.680
SplAdder Prec 0.963 0.844 0.734 0.513 0.957 0.857 0.733 0.450
  Rec 0.822 0.927 0.899 0.552 0.855 0.947 0.936 0.531
  FM 0.887 0.884 0.808 0.532 0.903 0.900 0.822 0.487
rMATS Prec 0.995 1 1 0.976 0.996 1 1 0.976
  Rec 0.905 0.685 0.755 0.830 0.905 0.685 0.755 0.830
  FM 0.948 0.813 0.860 0.897 0.949 0.813 0.860 0.897
SUPPA2 Prec 1 0.880 0.754 0.976 1 0.880 0.754 0.976
  Rec 0.894 1 1 0.830 0.894 1 1 0.830
  FM 0.944 0.936 0.860 0.897 0.944 0.936 0.860 0.897
  1. Results obtained by ASGAL, SplAdder and rMATS (for which we used STAR to compute the alignments), and SUPPA2 (for which we used Salmon to obtain the transcript quantification) are reported. Precision (Prec), Recall (Rec), and F-Measure (FM) achieved on the simulated datasets in detecting annotated alternative splicing events: exon skipping (ES), alternative acceptor site (A3), alternative donor site (A5), and intron retention (IR)