Table 1.
Runtime in CPU hours for each program on simulation 1 and 2
| Program | Simulation 1 | Simulation 2 |
|---|---|---|
| ContextMap Bwt1 | 11.67 | 11.02 |
| ContextMap Bwt2 (k=3, default) | 16.47 | 15.58 |
| ContextMap Bwt2 (k=10) | 24.98 | 24.55 |
| ContextMap Bwa | 11.58 | 14.00 |
| ContextMap Bwa ann | 11.92 | 14.15 |
| MapSplice 2 | 31.43 | 28.62 |
| STAR 1-pass | 0.82 | 1.28 |
| STAR 1-pass ann | 1.05 | 1.58 |
| STAR 2-pass | 9.60 | 10.28 |
| STAR 2-pass ann | 9.57 | 10.80 |
| TopHat1 | 20.1 | 28.43 |
| TopHat1 ann | 20.53 | 29.03 |
| TopHat2 | 25.17 | 27.23 |
| TopHat2 ann | 34.32 | 39.68 |
| GSNAP | 147.73 | 128.15 |
| GSNAP ann | 160.78 | 140.27 |
All methods were run using 8 cores on the same machines and with the same parameter settings as in the RGASP evaluation [17]. ContextMap 2 with Bowtie 2 was run with the maximum number of alignments reported per read (k) set to 3 (default setting used for evaluating mapping quality) and 10, respectively. Runtime of STAR 2-pass includes the time required for running STAR 1-pass, indexing the genome with splice sites found in the first STAR run and re-running STAR.