Skip to main content

Table 1 The sensitivity and false discovery rate (FDR) of the different SNV detection algorithms with different simulation setups

From: Pysim-sv: a package for simulating structural variation data with GC-biases

Purity Subclone1 Subclone2 GC-bias Methods Sensitivity FDR
Purity =1 1 0 Yes GATK 0.97679 1.13×10−4
     Varscan 0.97892 2.36×10−3
    No GATK 0.97826 9.20×10−5
     Varscan 0.98134 2.05×10−3
Purity =0.8 0.4 0.4 Yes GATK 0.90253 1.93×10−3
     Varscan 0.92501 5.36×10−3
    No GATK 0.90853 8.32×10−4
     Varscan 0.93001 5.05×10−3
Purity =0.5 0.4 0.1 Yes GATK 0.81253 1.09×10−2
     Varscan 0.84501 5.34×10−2
    No GATK 0.81853 2.11×10−3
     Varscan 0.85001 4.51×10−2
  1. Note that the purity column represents the proportion of “tumor” cells in the simulated data. Subcolone1 and Subcolone2 columns represent the proportions of subcolone 1 and 2 in the simulated data, respectively. When purity is 1 and subclone1 is 1, it means that all data are from subclone1 and it essentially like a sequencing data from a normal genome