Exome sequencing analysis problem

[albertyu]

Hello All,

I am new in next generation sequencing analysis. I am working on a vcf file which performed by GATK tools. I have a question about GT (genotype) for ChrY and ChrM. the data looks lke:
GT : GQ : DP : PL 0/1:8.82:96:4088,0,9
GT : GQ : DP : PL 1/1:38.92:1:390,38,0
As my understanding, GT means genotype. Since there is only one copy for chrY or chrM, why the data showed 0/1 (heterozygous) or 1/1 (homozygous)? Or do I misunderstand the meaning of "GT" ?

Thanks,
Albert

[Bukowski]

Well for mitochondria it could be :

http://en.wikipedia.org/wiki/Heteroplasmy

As for the Y - are they in pseudoautosomal regions? Or perhaps badly genotyped indels?

[albertyu]

Well for mitochondria it could be :

http://en.wikipedia.org/wiki/Heteroplasmy

As for the Y - are they in pseudoautosomal regions? Or perhaps badly genotyped indels?

Thanks for your reply. I checked the possibility of the pseudoautosomal genes. It seems that most of variants are not at the pseudoautosomal PAR1 and PAR2 genes location. And many of them are substitutions not indels. I would think it's because mis-mapping of BWA. ???

Actually I originally thought this is a popular problem in Exome sequencing data since I saw several whole exome sequencing vcf data containing chrY with GT:1/0 or 1/1.

[Bukowski]

Thanks for your reply. I checked the possibility of the pseudoautosomal genes. It seems that most of variants are not at the pseudoautosomal PAR1 and PAR2 genes location. And many of them are substitutions not indels. I would think it's because mis-mapping of BWA. ???

Actually I originally thought this is a popular problem in Exome sequencing data since I saw several whole exome sequencing vcf data containing chrY with GT:1/0 or 1/1.

I just had a look at some exome data, and I think there is definitely some mismapping involved as most females in my trios a small number of chrY variations. As I suspected though in my data, most of the het calls are small indels rather than SNPs.

I don't see what is wrong with a 1/1 call though, in Y that should just mean 'base is different to reference' surely?

[albertyu]

I just had a look at some exome data, and I think there is definitely some mismapping involved as most females in my trios a small number of chrY variations. As I suspected though in my data, most of the het calls are small indels rather than SNPs.

I don't see what is wrong with a 1/1 call though, in Y that should just mean 'base is different to reference' surely?

Indeed, all my female samples have chrY variations.
From my data performed by GATK, it separated substitutions and indels. So I originally thought those are all indels and finally I realized there are a lot of SNPs in the middle of vcf files.
If they are mostly mismapping, I am afraid to use these data. So I would filter out those data at chrY.

[nexgengirl]

Check out section 6 on this page:

http://gatkforums.broadinstitute.org/discussion/1237/using-the-unified-genotyper

It explains how the genotypes on the sex chromosomes are called in GATK

[AJERYC]

[QUOTE=albertyu;89712]Indeed, all my female samples have chrY variations.
From my data performed by GATK, it separated substitutions and indels. So I originally thought those are all indels and finally I realized there are a lot of SNPs in the middle of vcf files.
If they are mostly mismapping, I am afraid to use these data. So I would filter out those data at chrY.[/QUOTEI

In any exome sequencing you will get some "false SNP" that have different sources: sequencing errors, aligning pseudogenes, ... That is why in an exome sequencing you can see chromosome Y in women or heterozygous SNPs in men. You can eliminate these "errors" by increasing your level of filtering in the alignement process but then the price you pay is with a tougher filtering is that you can lost true information too.