Table 3. Effect of the strict clustering mode on the clustering properties when varying the distance threshold.
Analyses were run on 63,863 RefSeq Bacteria using the Jaccard Index and the direct strategy (JI-d) with six different distance thresholds (from 0.8 to 0.9). All pack sizes were 200. RI, Redundancy Index (# groups / # phyla). This table has to be compared to the upper-left quarter of Table 2.
| Thresholds | 0.80 | 0.82 | 0.84 | 0.86 | 0.88 | 0.90 |
|---|---|---|---|---|---|---|
| RI | 66 | 49 | 37 | 24 | 15 | 8 |
| # phyla | 34 | 33 | 28 | 26 | 20 | 14 |
| # groups | 2231 | 1609 | 1035 | 614 | 300 | 112 |
| - pure groups | 2220 | 1592 | 1021 | 598 | 283 | 104 |
| – singletons | 1289 | 875 | 551 | 328 | 149 | 52 |
| - mixed groups | 11 | 17 | 14 | 16 | 17 | 8 |
| – paraphyletic | 1 | 0 | 0 | 0 | 2 | 0 |
| – super phyla | 10 | 16 | 10 | 10 | 11 | 4 |
| – polyphyletic | 0 | 1 | 4 | 6 | 4 | 4 |