Skip to main content
. 2021 Aug 2;21:886. doi: 10.1186/s12885-021-08635-5

Fig. 3.

Fig. 3

Performance comparison between the multi-modal approach and ALLSorts. Accuracy comparison for BCP-ALL patients harboring entity-defining fusion transcripts or abnormalities in chromosome number (a) and the number of false positive calls (b). c Overlap in label class for different confidence levels (low: 50–80% probability; medium: 80–90% probability; high: > 90% probability). d Breakdown of the identified BCP-ALL ‘other’ cases by both approaches. Conf: confidence; FN: false negative; TP: true positive