Table 4.
Aligner performance in mapping junction reads and reporting exon–exon junctions. Results are based on simulated data
| Junctions | Junction reads | |||||
|---|---|---|---|---|---|---|
| Workflow | Recall | Prec | F1 | Recall | Prec | F1 |
| subjunc-F | 99.80 | 99.53 | 99.66 | 95.51 | 98.18 | 96.82 |
| subjunc-G | 99.82 | 99.43 | 99.63 | 95.50 | 98.16 | 96.81 |
| STAR | 98.48 | 99.87 | 99.17 | 89.87 | 98.06 | 93.78 |
| TopHat2 | 99.12 | 99.91 | 99.51 | 90.15 | 98.57 | 94.17 |
Column ‘Recall’ gives the percentage of correctly called junctions (or junction reads) out of all junctions (or junction reads) generated in the simulated dataset. Column ‘Precision’ gives the percentage of correctly called junctions (or junction reads) out of all reported junctions (or junction reads). Column ‘F1’ gives the F1 score that is the harmonic mean of precision and recall.