Table 5.
Quality validation of programs for LTR retrotransposon prediction on the genome of D. melanogaster
Program used | LTR_STRUC | LTR_seq | LTR_Rho | LTR_FINDER | LTR harvest | LTR harvest | LTR_seq | LTR_Rho | LTR_FINDER | LTR harvest |
Run-number | 1 | 2 | 3 | 4-1 | 5-1 | 5-2 | 6 | 7 | 8 | 9 |
Parameter set | default* | default | default | default | default | default + clustering | see Tab. 1 | see Tab.1 | see Tab. 1 | see Tab.1 + clustering |
Index files contruction [s] | - | - | - | - | 138 | 138 | - | - | - | 138 |
Run-time [s] | 4380 | 24120 | 2286 | 1209 | 25 | 198** | 1380 | 1709 | 320 | 170** |
Annotations 'full length' | 304 | 304 | 304 | 304 | 304 | 304 | 304 | 304 | 304 | 304 |
Predictions | 310 | 188 | 417 | 395 | 723 | 490 | 160 | 398 | 204 | 411 |
Sensitivity | 37.5% | 36.8% | 94.7% | 74.3% | 94.7% | 97.4% | 35.2% | 96.1% | 52.0% | 97.7% |
Specificity | 36.8% | 59.6% | 69.1% | 57.2% | 40.4% | 60.4% | 66.9% | 73.4% | 77.5% | 72.3% |
Annotations 'all' | 682 | 682 | 682 | 682 | 682 | 682 | 682 | 682 | 682 | 682 |
Predictions | 310 | 188 | 417 | 395 | 723 | 490 | 160 | 398 | 204 | 411 |
Sensitivity | 20.1% | 22.0% | 54.0% | 45.9% | 58.9% | 57.9% | 19.8% | 53.5% | 28.7% | 56.3% |
Specificity | 44.2% | 79.8% | 88.2% | 90.3% | 55.6% | 80.6% | 84.4% | 91.7% | 96.1% | 93.4% |
* = parameters are not adjustable.
** = run-time LTRharvest + clustering with Vmatch.
Details on parameter settings and exact numbers of predictions are shown in Table C of the Additional file 1. Sensitivity and specificity values were calculated against the D. melanogaster annotation counting all TPs and hTPs as true positives.