Skip to main content
Fig. 3 | BMC Genomics

Fig. 3

From: Improved detection of clinically relevant fusion transcripts in cancer by machine learning classification

Fig. 3

Characteristics of validated fusions and classifier performance for prediction of true positive fusion transcripts. A Compared to non-validated fusion transcripts (n= 861593), validated fusion transcripts (n=4237) had more supporting read pairs in RNA-Seq data, B longer anchor length for RNA-Seq reads mapped to the fusion junction, C a higher fraction of intrachromosomal fusions, and D lower average expression. E Precision-recall curve for classifier performance in prediction of true positive fusion transcripts for an independent cohort of 249 breast tumors. F Top predictive features color-coded by source of information

Back to article page