Do I need to use hg19 or hg38 reference for performing variant calling analysis and RNAseq analysis?

Question

[bioinforesearchquestions]

Dear All,

I need some help on selecting the reference fasta for human genome. Do I need to use hg19 or hg38 reference for performing variant calling analysis and RNAseq analysis?

[mastal511]

the link that I gave above leads to directories hg38/hg38bundle, so, yes.

Answer 1

[Vivek]

There is no wrong option here. I believe most of the annotation information has been made available for hg38, so you should be good to go with the newer version.

[bioinforesearchquestions]

Thanks Vivek.

1) I am using GATK for variant calling analysis. Where can I download the hg38 version of reference for gate_bundle? 2) For RNASeq analysis, I downloaded the hg38 version of reference from genome's Illumina website.

Do you suggest me any other standard website for downloading hg38?

[Vivek]

As of now the GATK bundle does not have a hg38 version. So if you want to avoid that hassle, best to go with hg19.

[bioinforesearchquestions]

Thanks Vivek for help. Then I will go with hg19 version itself.

Answer 2

[mastal511]

You will save yourself a lot of hassle by downloading the genome from GATK, otherwise you will keep getting errors because the chromosomes are not ordered in 'karyotypic' order the way GATK uses them.

https://www.broadinstitute.org/gatk/guide/article.php?id=1215

[bioinforesearchquestions]

Dear Mastal,

I have GATK;s hg19 bundle resources. Does GATK bundle provides hg38 resources or not?

Answer 3

[H.Hasani]

Hi,

as Vivek already said, both options are valid, the question is how does the downstream analysis look like, you might face tools, that can not tolerate one of those...just keep that in mind!

[bioinforesearchquestions]

I am going to do variant calling analysis using GATK and annovar. I might compare my snps with dbsnp, 1000genomeproject, hap map etc..

Answer 4

[Ashley]

Hi, if the software which the next step's need is hg19, I perfer you to hg19.