I ran verifyBamID to check contamination and sample swapping for 2 projects. All data is from WES.
bam file is from tumor sample, and germline mutations list as vcf file is from all blood samples. Ran it on --best mode.
(1) First Project: the last 2 columns are for FREEMIX and CHIPMIX:
Tumor MatchingBlood BestMatchBlood FREEMIX CHIPMIX
T99633 B99633 B99633 0.04319 0.96899
T136801 B136801 B134340 0.05034 0.97350
T96096 B96096 B134340 0.04670 0.97649
T9 B174 B134340 0.04996 0.97661
T78254 B78254 B134340 0.03918 0.98401
T127554 B127554 B134340 0.03603 0.98438
FREEMIS is larger than the threshold that verifyBamID document recommends (0.02), but still quite acceptable. But CHIPMIX is way too large. Since the genotype data is not from the same tumor sample, but from matching blood sample. So I suppose it is expected to be larger than what the document suggest. But how large is acceptable? Would 0.50 acceptable? The result above means that these tumor samples identified the correct matching blood sample as their pairs, but the matching tumor and blood samples are having very poor agreement on germline mutations. So we should conclude that these are not really matching samples. Am I right?
(2) Second project:
Tumor Pair_Blood BestMatch_Blood FREEMIX CHIPMIX
N045 B68756 B68756 0.37483 0.36563
FREEMIX is too high. This means the sample is highly contaminated, and we cannot use this sample, right?
