Hi all,
We have implemented the last software available of Illumina in our Myseq sequencer and I am comparing the data generated between the previous version and the new version changing some parameters. I have one VCF (actually this is 16 samples combined) from the previous version and 4 VCF files after running the most recent software turning off and on two parameters. Obviously, all data come from the same sequencer output.
As a learning activity ( I am a trainee) I have been asked to get the sensitivity and specificity of the new version assuming that the variants that passed all filters and got quality equal to 100 are the positives in these tests.
Having this I think the most appropriate statistical test is McNemar's test.
Does this analysis make sense and are the FILTER and QUAL attributes correct to analyse sensitivity and specificity?
