benchmarking_toolkit

Briefly, the Benchmarking Toolkit first maps genes to gencode symbols; the script “map_gene_symbols_to_gencode_FID.pl” adds lists of gene alias symbols which map to each of the fusion partners. Following this, FusionAnnotator adds metadata to the fusion predictions (described in further detail below). A “fusion_preds_to_TP_FP_FN_FID.pl” script scores entries as true positive, false positive and false negative. Finally, the number of entries in each of these three categories is counted by a custom script we have developed, which then provides this information as an output count file. These scripts are identical to those published by Haas et al (2019) except for the addition of a unique FID to the calls. The only difference beween BH's scripts and my version of them is the inclusion of FID as a column in the files they produce

Link to BH's Wiki for the benchmarking toolkit:

https://github.com/fusiontranscripts/FusionBenchmarking/wiki