Dear all, I'm using gatk-126.96.36.199 to compare variants among two WGS from MZ Twins. As a summary I can say that the 97% of all variants are common in both samples as expected. But there are a big number of discordant variants among both samples annotated in dbSNP, which is quite impossible to believe, as both samples come form the same zygote.
Among the discordant variants I have several examples of clear variants called in one sample but not in the other, although having similar values (scores, DP, GQ etc etc). I've tried to find an explanation but I can't.
For example, in chr1 position 125178240 of one sample I see a variant with a score of 37014, DP is 4199 and GQ 99. According to IGV, in this position the the 76% are from the reference (G) and ~ 24% from the alt (A). If I check this position in the sample where the variant is missing, I see similar values:
I tried to use varscan2 to see how this position is detected. According to the result, this is a germline variant with a frequency of 23.58% in one sample and 24.44% in the other.
I have more example quite similar to this.
I've used BWA as aligner, then I`ve used gatk MarkDuplicatesSpark and finally the HaplotypeCaller using default parameters except: --native-pair-hmm-threads 20
So ... Has any of you an explanation of this behaviour ?
Please sign in to leave a comment.