Table 4.
Validation of annoFuse prioritization using PPTC PDX fusion calls
| Histology | PPTC STAR-Fusion/FusionCatcher/SOAPFuse/ deFuse (n detected) | PPTC STAR-Fusion/Arriba (n detected) | annoFuse STAR-Fusion/Arriba (n retained) | % Retained with annoFuse |
|---|---|---|---|---|
| ALL | 117 | 75 | 72 | 96 |
| CNS Embryonal | 4 | 3 | 3 | 100 |
| Ependymoma | 2 | 0 | 0 | NA |
| Ewing Sarcoma | 11 | 10 | 10 | 100 |
| Glioblastoma | 1 | 1 | 0 | 0 |
| Osteosarcoma | 18 | 15 | 15 | 100 |
| Other Brain | 1 | 1 | 1 | 100 |
| Other Sarcoma | 4 | 3 | 3 | 100 |
| Rhabdomyosarcoma | 7 | 6 | 6 | 100 |
| Wilms | 1 | 0 | 0 | NA |
| Total | 166 | 114 | 110 | 96 |
| ALL truth set | 27 | 23 | 23 | 100 |
Retention of high-confidence, putative oncogenic calls averaged 96% across the entire PPTC PDX dataset and was 100% for the ALL truth set (ALL = acute lymphoblastic leukemia). Column 1 = PPTC histology, Column 2 = fusion calls from STAR-Fusion, FusionCatcher, deFuse, and SOAPFuse which were filtered and reported as high-confidence in the PPTC dataset, Column 3 = PPTC reported fusions detected from STAR-Fusion and Arriba, Column 4 = Fusions retained following annoFuse filtering, Column 5 = Percent of fusions retained after applying annoFuse