search

stat-lab / EvalSVcallers

Evaluate the performances (precision and recall) of structural variation (SV) callers
32 stars 13 forks source link

The script "convert_SV_callers_vcf.pl" discarded a significant number of CNVnator results. #19

Open jingydz opened 8 months ago

jingydz commented 8 months ago

I obtained 5909 results using the CNVnatorv0.3.3 software on the 30x BAM file of NA12878 from the 1000 Genomes Project. However, after converting the results with the "convert_SV_callers_vcf.pl" script, I obtained only 797 results. Is this normal?

Why the script "convert_SV_callers_vcf.pl" discarded a significant number of CNVnator results?

5909 NA12878.cnvnator.cnv $ ./EvalSVcallers/scripts/convert_SV_callers_vcf.pl --tool CNVnator NA12878.cnvnator.cnv |wc -l 797

\$ head -n 4 NA12878.cnvnator.cnv deletion chr1:1-10000 10000 0 1.59373e-11 2.21635e-296 1.99216e-11 2.33409e-235 -1 deletion chr1:44501-48000 3500 0.587305 71.8268 0.00397571 39604.1 23755.7 0.966805 deletion chr1:51001-65000 14000 0.673617 3.89264e-05 211907 3.6223e-05 824925 0.401072 deletion chr1:66001-93000 27000 0.292106 5.90269e-12 8.33278e-69 6.3749e-12 4.61625e-63 0.758621

\$ ./EvalSVcallers/scripts/convert_SV_callers_vcf.pl --tool CNVnator NA12878.cnvnator.cnv |head -n 3 chr1 66001 DEL . . . PASS SVTYPE=DEL;SVLEN=27000;RDRATIO=0.292106;READS=7;GT=1/0 chr1 142001 DEL . . . PASS SVTYPE=DEL;SVLEN=14000;RDRATIO=0.549935;READS=3;GT=1/0 chr1 203001 DEL . . . PASS SVTYPE=DEL;SVLEN=58000;RDRATIO=0.0887561;READS=4;GT=1/1

stat-lab commented 8 months ago

The sv number of your final output is normal, but the number of your CNVnator output is too high (I obtained about 1,400 SVs for NA12878). Please see the Supplemental Methods of our paper for CNVnator command options and filtering conditions encoded in convert_SV_callers_vcf.pl/convert_CNVnator_vcf.pl.