Skip to main content

Table 2 Results of the simulation study for 9 taxa

From: An invariants-based method for efficient identification of hybrid species from large-scale genomic data

  Coalescent Independent Sites Multi-locus Data
  Shallow Hybridization Deep Hybridization Shallow Hybridization Deep Hybridization
γ False True True False True True False True True False True True
  Pos. Pos. Sets Pos. Pos. Sets Pos. Pos. Sets Pos. Pos. Sets
0 0.064 0.048 0.216 0.224
0.1 0.024 0.384 0.192 0.024 0.664 0.520 0.176 0.424 0.312 0.088 0.719 0.552
0.2 0.032 0.968 0.920 0.000 0.952 0.944 0.000 0.968 0.864 0.000 1.000 0.896
0.3 0.032 0.976 0.976 0.000 0.976 0.976 0.000 1.000 0.968 0.000 1.000 1.000
0.4 0.008 0.976 0.144 0.000 1.000 0.448 0.000 1.000 0.248 0.000 1.000 1.000
0.5 0.016 0.960 0.000 0.000 0.952 0.000 0.000 1.000 0.000 0.000 1.000 1.000
  1. The columns labeled “False Pos.” refer to the proportion of data sets for which a triplet of taxa were incorrectly identified as involving a hybridization event (false positives); the columns labeled “True Pos.” refer to the proportion of data sets for which the correct triplet of taxa involving the hybridization event was identified and the hybrid taxon was correctly identified (true positives); and the columns labeled “True Sets” refer to the proportion of data sets for which the correct triplet of taxa was identified but the hybrid taxa was specified incorrectly. It is possible that both the correct triplet with the hybrid correctly specified and the correct triplet with the hybrid misspecified are identified as statistically significant in the analysis. We tally these separately because in the case of empirical data it would be ambiguous as to which is the hybrid taxon. For our simulated data, all data sets for which the true set was significant also had the triplet with the correct hybrid assignment found to be significant, and thus this proportion is always a fraction of the proportion of true positives