Hi.
I'm evaluating a new NGS prep kit and want to calculate the robustness of the method by comparing the variant call from a samples which have been prepped twice.
I've taken the approach of using rtg vcfeval, and then my calculation has has just been to take the true-pos-call divided by tru-pos call + false-pos + false-neg from the summary.txt file.
I've been looking at using the --squash-ploidy option as well, but I'm not sure about it since I am using germline human whole genome samples.
I've also seen people just using the F-measure as the robustness score, but I'm not sure if I truly understand how that is applicable.
Is there anything wrong with the approach here? Should I be using --squash-ploidy for these samples? And how should one interpret the F-measure here?
Cheers!
I'm evaluating a new NGS prep kit and want to calculate the robustness of the method by comparing the variant call from a samples which have been prepped twice.
I've taken the approach of using rtg vcfeval, and then my calculation has has just been to take the true-pos-call divided by tru-pos call + false-pos + false-neg from the summary.txt file.
I've been looking at using the --squash-ploidy option as well, but I'm not sure about it since I am using germline human whole genome samples.
I've also seen people just using the F-measure as the robustness score, but I'm not sure if I truly understand how that is applicable.
Is there anything wrong with the approach here? Should I be using --squash-ploidy for these samples? And how should one interpret the F-measure here?
Cheers!