Table 8. Inversion detection task.
Test Case | Chrom. | Inversion | Clusters | PC-SNP | Cluster-SNP |
---|---|---|---|---|---|
Negative | D. mel. 3L | None | 1 | No | No |
Negative | 34 An gam. and col. 3L | None | 4 | No | No |
Negative | 150 An gam. and col. 3L | None | 2 | No | No |
Single | D. mel. 2L | In(2L)t | 3 | Yes (PC 1) | Yes |
Single | D. mel. 2R | In(2R)NS | 3 | Yes (PC 1) | Yes |
Single | D. mel. 3R | In(3R)Mo | 3 | Yes (PC 1)* | Yes* |
Single | In(3R)p | 3 | Yes (PC 1)* | Yes* | |
Single | In(3R)K | 3 | Yes (PC 1)* | Yes* | |
Multiple | 150 An. gam. and col. 2L | 2La | 3 | Yes (PC 2) | No |
Multiple | 81 An. gam. 2L | 2La | 2 | Yes (PC 1) | Yes |
Multiple | 34 An. gam. and col. 2L | 2La | 4 | Yes (PC 1) | Yes |
We compared inversions detected by the three methods to the known inversion karyotypes for these data sets taken from the original papers describing the data [17, 37–39]. If an inversion was present with no population structure, three clusters corresponding to three possible genotypes (which may not all be present) would be expected.
*Multiple, mutually-exclusive inversions were detected as a single inversion by our methods.