Table 3. DNA Barcoding statistics and performances.
| IDa | Setsb | Excludedc | Nd | eOTe | CEeOTf | NNg | Ah | Pi |
|---|---|---|---|---|---|---|---|---|
| ds1 (107) | L3 | 24 (8) [1,20] | 1.4–4.8 | 26 (0, 26) | 86T, 21F | 76 | 100 | |
| ds2 (95) | ds2 ⊂ ds1 | L3, sng | 13, 7 [2,20] | 0.7–0.8 | 25 (6,19/0,25) | 84T, 11F | 74 | 92 |
| ds3 (90) | ds3 ⊂ ds2 | L3, sng, D. cinerella | 12, 7 [2,20] | 0.8,1.0 | 9 (0,9/1,8) | 88T, 2F | 90 | 99/100 |
| ds4 (89) | ds4 ⊂ ds3 | L3, sng, D. cinerella, hybrid | 12, 7 [2,20] | 1.0–5.3 | 0 (0,0) | 89T | 100 | 100 |
| ds5 (75) | ds5 ⊂ ds2 | L3, sng, D.tonsa | 12, 6 [2,14] | 0.8 | 3 (0,3) | 73T, 2F | 96 | 100 |
| ds6 (74) | ds6 ⊂ ds5 | L3, sng, D. tonsa, hybrid | 12, 6 [2,14] | 0.8–5.2 | 0 (0,0) | 74T | 100 | 100 |
a Identifier of each analyzed datasets.
b Logical relation among datasets.
c List of excluded cox1 sequences respect to the 112 obtained; L3: larvae at 3rd developmental stage; sng: singletons; hybrid: hybrid specimen between D. vaillanti and D. tonsa.
d Number of morphospecies included in the dataset; within brackets the average number of specimens per species; within square brackets the minimum and maximum number of specimens per species.
e Estimated optimal threshold: nucleotide distance or range of distances, expressed as percentage, that minimize the function Fx = min∑ (FP+ FN).
f CEeOT: cumulative error at eOT, within brackets are reported the number of FP and FN.
g Near neighbor defined as Maier et al. (2006), number of tested sequences with as closest individual a conspecific (true, T) or a non conspecific specimens (false, F).
h Accuracy calculated as follow: A = (TP+TN)/n° sequences; values are expressed as percentage.
i Precision calculated as follow: P = TP/(TP+FP); values are expressed as percentage. FP: false positive identification corresponding to type I errors; FN: false negative corresponding to type II errors; TP: true positive; TN: true negative.