Table 5.
Compression of GAhum Using SlimGene
| Fragments+ alignment (GB) | Q-values (GB) | total (GB) | Execution time (hr) | |
|---|---|---|---|---|
| Uncompressed | 124.7 | 103.4 | 228.1 | N/A |
| gzip (in isolation) | 15.83 | 49.92 | 65.75 | N/A |
| bzip2 (in isolation) | 17.9 | 46.49 | 64.39 | 10.79 |
| SlimGene | 3.2 | 42.23 | 45.43 | 7.38 |
| SlimGene+bzip2 | 3.04 | 42.34 | 45.38 | 7.38 |
| SlimGene+lossy Q-values (b = 3) | 3.2 | 26 | 29.8 | 7.38 |
| SlimGene+lossy Q-values (b = 1) | 3.2 | 13.5 | 16.7 | 7.38 |
Using a loss-less Q-value compression, we reduce the size by 5×. A lossy Q-value quantization results in a further 3× compression, with minimal effect on downstream applications.