Table 3.
Computational benchmarking of svtools subcommands
| Num. samples |
10 |
100 |
1000 |
|||
|---|---|---|---|---|---|---|
| Program | Wall (m) | RAM (MB) | Wall (m) | RAM (MB) | Wall (m) | RAM (MB) |
| lsort | 0.129 | 5.964 | 1.117 | 1696.008 | 16.788 | 3402.480 |
| lmerge | 2.108 | 87.791 | 18.708 | 258.402 | 193.346 | 2032.114 |
| genotype | 13.425 | 2008.828 | 31.725 | 1222.536 | 61.413 | 1255.593 |
| copynumber | 0.225 | NA | 0.333 | NA | 0.533 | NA |
| vcfpaste | 0.088 | NA | 1.379 | 75.660 | 79.083 | 181.845 |
| afreq | 0.096 | NA | 0.908 | 77.701 | 20.192 | 97.713 |
| vcftobedpe | 0.083 | NA | 0.183 | 3.277 | 0.933 | 70.851 |
| bedpesort | 0.079 | 17.363 | 0.183 | NA | 0.892 | 70.904 |
| prune | 0.179 | 18.790 | 0.404 | 61.667 | 1.388 | 171.728 |
| bedpetovcf | 0.079 | 17.368 | 0.183 | 34.794 | 0.879 | 71.003 |
| vcfsort | 0.033 | NA | 0.100 | 0.594 | 1.508 | 900.887 |
| classify | 6.938 | 530.722 | 8.250 | 526.480 | 25.621 | 680.900 |
Note: For three different size cohorts, each tool was run (n = 4; n = 3 for the 100 sample bedpetovcf) to generate mean wall clock time and RAM utilization. For the genotype and copynumber commands, benchmarking was performed on a single, representative sample within the cohort of median file size. All other commands were evaluated on the entire dataset. Some benchmarking runs finished before LSF was able to gather memory usage metrics and these are reported as NA.