Table 2. Summary of cMonkey2 module construction on M. tuberculosis expression data set, including varying combinations of additional prior data (ChIP-seq and TFOE; see Methods) via the set-enrichment scoring function.
| Expression | ChIP-Seq(a) | TFOE(a) | Both(a) | |||||
|---|---|---|---|---|---|---|---|---|
| Motif | No Motif | Motif | No Motif | Motif | No Motif | Motif | No Motif | |
| mean residual | 0.53 ± 0.01 | 0.49 ± 0.01 | 0.53 ± 0.01 | 0.49 ± 0.01 | 0.53 ± 0.01 | 0.50 ± 0.01 | 0.53 ± 0.01 | 0.50 ± 0.01 |
| mean motif log10 p-val. | -9.14 ± 0.07 | – | -9.21 ± 0.10 | – | -9.19 ± 0.08 | – | -9.31 ± 0.12 | – |
| clusters w. motifE ≤ 1 | 564.8 ± 7.6 | – | 565.3 ± 7.2 | – | 563.0 ± 2.9 | – | 571.5 ± 3.4 | – |
| ChIP-seq | ||||||||
| clusters signif.(1) | 349.5 ± 5.8 | 187.3 ± 7.7 | 397.5 ± 8.5 | 418.1 ± 30.4 | 349.9 ± 11.2 | 182.8 ± 9.9 | 395.3 ± 8.9 | 419.2 ± 11.9 |
| TFs signif.(2) | 142.0 ± 0.0 | 133.4 ± 3.5 | 142.0 ± 0.0 | 141.5 ± 1.0 | 142.0 ± 0.0 | 129.9 ± 4.8 | 142.0 ± 0.0 | 142.0 ± 0.0 |
| unique clusters signif.(1) | 95.7 ± 3.2 | 87.1 ± 3.9 | 97.6 ± 4.3 | 116.2 ± 2.5 | 93.0 ± 3.8 | 88.1 ± 4.0 | 98.3 ± 3.9 | 118.4 ± 4.3 |
| unique TFs signif.(2) | 135.6 ± 1.8 | 134.7 ± 2.9 | 137.0 ± 1.3 | 136.5 ± 1.5 | 136.3 ± 1.6 | 136.2 ± 1.2 | 136.5 ± 1.7 | 135.8 ± 3.3 |
| TFOE | ||||||||
| clusters signif.(1) | 346.3 ± 7.8 | 249.1 ± 9.7 | 347.8 ± 12.5 | 255.2 ± 10.3 | 402.2 ± 11.3 | 485.5 ± 15.8 | 413.1 ± 14.9 | 491.1 ± 7.2 |
| TFs signif.(3) | 198.8 ± 1.8 | 179.0 ± 3.9 | 200.0 ± 1.9 | 181.4 ± 5.6 | 201.6 ± 1.7 | 199.5 ± 1.8 | 202.8 ± 1.0 | 199.8 ± 5.4 |
| unique clusters signif.(1) | 138.9 ± 3.8 | 136.2 ± 5.3 | 135.6 ± 2.4 | 137.2 ± 5.7 | 144.4 ± 3.5 | 174.9 ± 4.6 | 146.3 ± 8.2 | 171.9 ± 4.2 |
| unique TFs signif.(3) | 191.7 ± 2.7 | 189.2 ± 3.4 | 191.6 ± 2.6 | 188.4 ± 3.2 | 189.6 ± 2.9 | 190.8 ± 2.3 | 191.0 ± 5.0 | 190.7 ± 2.5 |
Shown are statistics regarding recapitulation of TF target gene sets in the ChIP-seq and TFOE measurements, for cMonkey2 runs on data with inclusion of varying prior information. The values in the table rows labeled ‘clusters signif.’ (number of significant clusters) correspond to those in the bar chart of Figure S3. The rows labeled ‘Unique clusters/TFs signif.’ denote clusters/TFs with a single unique match to a TF/cluster, respectively, and respectively represent precision and recall.
Notes: (a) All runs included expression data as well. (1) Out of a total of 600 clusters predicted. (2) Out of 142 TFs tested via ChIP-seq. (3) Out of 205 TFs tested via overexpression (TFOE).