3
$\begingroup$

I'm trying to find the HLA-B*15:01 variant in my DNA results, prompted from this research paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8142661/#:~:text=HLA%2DB*15%3A01%20is%20strongly%20associated%20with%20asymptomatic,mechanism%20underlying%20early%20viral%20clearance.

I have DNA test results from Dante Labs that contain the following files

"*.bam"
"*.bam.bai"
"*.cnv.vcf.gz"
"*.cnv.vcf.gz.tbi"
"*.filtered.indel.vcf.gz"
"*.filtered.indel.vcf.gz.tbi"
"*.filtered.snp.vcf.gz"
"*.filtered.snp.vcf.gz.tbi"
"*1_SA_L001_R1_001.fastq.gz"
"*SA_L001_R2_001.fastq.gz"

I tried loading this on the Genome Explorer on Sequencing.com, but it couldn't find HLA-B at all, only HLA-A and HLA-C.

Then I tried loading the VCF files in IGV. I was able to filter to Chromosome 6 and HLA-B, which resolved to position

chr6:31,319,652-31,326,956

The CNV VCF says no variables found for chr6:31,319,652-31,326,956 The filtered Indel VCF did produce some variants for this position, but I have no clue how to search deeper specifically for HLA-B*15:01. Here is a screenshot: enter image description here

The filtered SNP VCF produced alot more variants for this position, but again, I don't know how to go deeper than this.

I'm not really sure what to do from here?

I don't have data on what the reference was, so that is unknown to me. I will reach out to Dante Labs about that, is there some sort of standard that is supposed to be used, or is it different for every genome provider?

How do I actually perform the check in IGV assuming the reference did not have it and the VCF did. I also assume this VCF does not have this allele, but I frankly am lost into how to confirm it to the "last mile" through the VCF so to speak, or even just go straight to BAM

$\endgroup$
0

2 Answers 2

2
$\begingroup$

Typical SNP calling wouldn't give you good results on classical HLA genes. Ignore the VCF. Try a dedicated HLA genotyper on raw reads instead. I heard optitype is a good one for exome or whole-genome data. I haven't used it myself, though.

$\endgroup$
1
$\begingroup$

The specific question and the easiest question is the absence of HLA-B*15:01 from your sample and that is not unexpected at all.

The maximum frequency of this allele is 40% found in "China Beijing pop 2" , this however is really unusually high and was published by Guang Yang et al (2009, IJI 37, 205-212).

Mostly this allele is present at 5% for moderate frequency of occurrence for example Europe, such as in Germany here. Moreover, in some populations the frequency is very low <<1%, e.g. India or Israel.

So a sample genome at random the will not have the B*15:01.


The more complicated issue is the VCF and the reference genome. This is because the basis on whether an allele is present or absent is a comparison against the reference genome which will generate the VCF. Thus understanding the HLA-B alleles of the reference genome is really important. You need a reference genome which lacks HLA-B*15:01, therefore if your VCF also lacks the allele, its missing. However, if HLA-B*15:01 is present in the reference genome your VCF will not present the allele if its present.

The probability is that the allele is missing from both the reference and sample genome.


Comments

Easy way The easiest way is search the IGV for LT618832.1 (HLA-B*15:01 reference), IGV has a motif finder, here https://software.broadinstitute.org/software/igv/motif_finder

Hard way This is how I'd do it ..

  • convert the bam to fasta then
  • run a blast against Genbank reference LT618832.1

I'd section the bam file into the locus of interest (e.g. chromosome 6) e.g. ... samtools view ... ChrName: ... > chromo6.bam: you'd need to cross check how to do this last bit ... then via samtools >1.3,

samtools fasta chromosome6.bam > chromosome6.fasta

Finally you download a standalone version of blast, I assume its on -c bioconda ...

blastn -query HLAB1501.faa -db chromosome6.fa -out myhits.txt

From memory I don't think you need to makeblastdb for a fasta onto a fasta db.

You'd need to map the SNPs in introns parsing (the coordinates are in the Genbank reference), or optically against the Genbank reference onto the blast output.

You may want to cross check the HLA-B*15:01 against https://www.ncbi.nlm.nih.gov/nuccore/?term=HLA-B%2A15%3A01

[but be careful about the introns]

motif_finder seems easier and accommodates potential heterozygosity.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.