GT and GL fields in VCF file

Question

[ChIP]

Hi!

This might sound completely stupid, lazy and silliest question. But please help in understanding these acronyms of VCF file. I did had a look at the format pdf of VCF file and I got further confused. So, I have my data in a VCF file version 4.2:

chr1 10043 . T C 21 PASS 3 GT:DP:FT:GQ:GL 1/0:8:PASS:20:-2.0739,-0.00379451,-27.336
chr1    10055    .    T    G    4    PASS    3    GT:DP:FT:GQ:GL    0/0:13:PASS:31:-0.00031839,-3.6121,-54.5012
chr1    10105    .    A    C    7    PASS    3    GT:DP:FT:GQ:GL    1/0:45:PASS:13:-1.58548,-0.0193946,-153.729

Now, what does GT here represent for, it stand for Genotype and the different values that I have for GT, what does that represents and consequence can they have on a SNP call? Can I also determine if the mutation is one copy or in both copies of gene ? If yes how?

For GQ, I guess it is a phred score with -10log10p and higher the better?

The genome used is hg19.

Please help.

Thank you

Answer 1

[iraun]

If you read the VCF format specification pdf (http://samtools.github.io/hts-specs/VCFv4.1.pdf) you'll have the answer. Summarizing, GT represent the genotype, encoded as allele values separated by / or |. If the allele value is 0, means that it is equal to the reference allele (what is in REF field), if 1 mean that is equal to alternative (first allele listed in ALT), and if 2 it is equal to the second allele listed in ALT (if it exists).

So a SNP tagged with GT = 1/1 represent a SNP homozygous for the ALT allele (1/0 heterozygous, 0/0 homozygous for the reference).

For determining if the call is homozygous or heterozygous I suggest you to read this thread: How To Distinguish Heterozygotes And Homozygotes From Variants In Vcf Format?. It can be done according to different criteria.

[ChIP]

Hi,

Thank you for the answer, since I am making rules to identify homozygous and heterozygous calls in my data. I have following combinations in my GT field.

0/0
1/0
1/1
2/0
2/1
2/2
3/0
3/1
3/2
3/3

I know first is homozygous call and second is heterozygous call, but what about the others? Could you please help.

[biomagician]

Hi,

I have some GT that is ./. Do you know what this means? I am assuming no reads align to this position and we do not know what the genotype is there? Best, Thanks. C.

[Matteo Schiavinato]

./. means that there is not enough information. It depends on the thresholds you set when you call variants: say that you want at least coverage 10 to call a variant and you have only 8 reads all on one strand, this is not easy to judge by the algorithm so places there a ./. (or at least this is my understanding of this issue).

I remember it's written somewhere here: https://samtools.github.io/hts-specs/VCFv4.2.pdf

[rse]

Hi, sometimes i see 1/. or ./1 or 0/. or ./0 What does it mean?

[Matteo Schiavinato]

• 0 = reference sequence allele
• 1 = first variant allele
• 2 = second variant allele
• ...
• n = n variant allele

In your case you called variants for 3 different lines on 1 reference, thus you have numbers ranging from 0 to 3. Whenever the numbers are equal you have a homozygous call, whenever they're not you have a heterozygous call.

[GabrielMontenegro]

If the type is GT does it means it's only SNPs? How can I know that?

[Matteo Schiavinato]

If you called the variants, you will see which kind of variant that is by looking in the field number 8 of your VCF file (INFO field).

Answer 2

[Cytosine]

AFAIK the GT field tells you if a call for a given variant is a:

• 0 - reference call
• 1 - alternative call 1
• 2 - alternative call 2
• ...

In your case the first and the third variants are considered heterozygous, as both the reference allele and an alternative allele appear in their pileups. The second variant is considered a homozygous call, with only the reference allele present.

Yes, the GQ field is a phred score with higher values being better.

[ChIP]

Hi,

Thanks for explaining. one more thing, if GT is 2/0, then also the call is heterozygous? Since, I am making a small awk one liner to filter these homo or heterozygous calls can yuou please confirm these:

0/1, 1/1, 1/1, 1/0 and 0/0 -- hetero, homo,homo, hetero calls or can you suggest any other alternative rule to recognise the possible homo and hetero calls?

[Cytosine]

Sorry for the late reply.

Just about any combination of differing alleles means heterozigosity (0/1, 1/0, 1/2, 2/3, 3/1, 2/0,...)

I use this oneliner to separate homo- and hetero- zygous calls. Just switch the '==' for '!=' respectively.

grep -v "#" my-favorite-snps.vcf | awk '{print $NF}' | awk -F ":" '{print $1}' | awk -F "/" '$1==$2 {print}'

[jespinoz]

How can I add missing fields? My fields are: I16=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0;MQ0F=0;MQSB=1;SGB=-0.379885;RPB=1;MQB=1;BQB=1;DP=25;QS=1,0,0.5