Table 5. Comparison of RAM memory usage and CPU times.
| Program | FACS | MetaPhyler | CARMA | PhyloPythia | HiSeq | MiSeq |
|---|---|---|---|---|---|---|
| 269bp | 300bp | 265bp | 961bp | 92bp | 156bp | |
| CPU Runtime (minutes) | ||||||
| CARMA a | 290880 | 77340 | 74950 | 360107 | — | — |
| MEGAN a | 288020 | 72060 | 72010 | 351060 | — | — |
| MetaPhyler a | 10 | 20 | 2 | 28 | — | — |
| MG-RAST a | 60 | 10080 | 20160 | 12960 | — | — |
| LMAT kML | 36(60 b ) | 58 | 43 | 348 | — | — |
| LMAT kFull | 54(93 b ) | 213 | 38 | 772 | 15 | 33 |
| MiniKraken | — | 1.22 | 1.07 | 2.95 | 1.3 | 1.2 |
| CoMeta allDb | 41(76 b ) | 14 | 28 | 144 | — | — |
| CoMeta micDb (ph) | — | 9 | 14 | 35 | 8 | 9 |
| CoMeta micDb (ge) | — | — | — | 79 | 42 | 68 |
| Memory Usage (Megabytes of RAM) | ||||||
| CARMA a | 100 | 100 | 100 | 120 | — | — |
| MEGAN a | 1024 | 1024 | 1024 | 1410 | — | — |
| MetaPhyler a | 5734 | 5734 | 5734 | 5734 | — | — |
| MG-RAST a | — | — | — | — | — | — |
| LMAT kML | 17000(17284 b ) | 17019 | 2128 | 13311 | — | — |
| LMAT kFull | 9295(9481 b ) | 13247 | 13286 | 15092 | 5807 | 12392 |
| MiniKraken | — | 4098 | 3210 | 4100 | 1317 | 1449 |
| CoMeta allDb | 71260(71903 b ) | 70743 | 71313 | 69508 | — | — |
| CoMeta micDb | — | 19552 | 19320 | 19552 | 10297 | 17689 |
a—The results of the program are taken from the Bazinet–Cummings’ paper [56].
b—The results for FACS 269bp dataset, where reads with more than 50% of unknown nucleotides (Ns) are filtered out. The values outside the brackets are for the whole dataset.
FACS 269 bp, MetaPhyler 300 bp, and CARMA 265 bp datasets were classified to phylum level, whilst PhyloPythia 961 bp, HiSeq 92 bp, and MiSeq 156 bp datasets to genus level. In the table besides the times of classification to the genus level for CoMeta micDb (ge), the times of classification to earlier levels are shown—the phylum levels (ph).