Skip to main content
. 2021 Jan 14;49(6):e33. doi: 10.1093/nar/gkaa1237

Figure 4.

Figure 4.

(A, B) Sensitivity and number of calls for the callers on the HPV HIVID dataset. Although the reliable ground truth is only a small set of Sanger validated integrations, it is easy to see that the number of calls BatVI, ViFi and VIcaller predict is unrealistic, and it is most likely the results of tens of thousands of false positives. (C, D) The same analysis is repeated on detected HBV integrations in HBV HIVID dataset. For a subset of validated calls (marked as ‘Corrected’), SurVirus actually selects a better location compared to what was reported by Zhao et al. (21).