Reference and dbSNP incompatibility issue (MuTect2)

Question

[umn_bist]

When I try using MuTect2 (from GATK) I get this error

Is there a link to an (old) dbSNP that is compatible with UCSC's hg19 assembly?

EDIT: I cannot post the error message because Biostar is saying that it isn't in English... I used the dbSNP from NCBI ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/

00-All.vcf.gz

and I am using ucsc.hg19.fasta reference assembly

##### ERROR   dbsnp contigs = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X, Y, MT, GL000207.1, GL000226.1, GL000229.1, GL000231.1, GL000210.1, GL000239.1, GL000235.1, GL000201.1, GL000247.1, GL000245.1, GL000197.1, GL000203.1, GL000246.1, GL000249.1, GL000196.1, GL000248.1, GL000244.1, GL000238.1, GL000202.1, GL000234.1, GL000232.1, GL000206.1, GL000240.1, GL000236.1, GL000241.1, GL000243.1, GL000242.1, GL000230.1, GL000237.1, GL000233.1, GL000204.1, GL000198.1, GL000208.1, GL000191.1, GL000227.1, GL000228.1, GL000214.1, GL000221.1, GL000209.1, GL000218.1, GL000220.1, GL000213.1, GL000211.1, GL000199.1, GL000217.1, GL000216.1, GL000215.1, GL000205.1, GL000219.1, GL000224.1, GL000223.1, GL000195.1, GL000212.1, GL000222.1, GL000200.1, GL000193.1, GL000194.1, GL000225.1, GL000192.1, NC_007605]
##### ERROR   reference contigs = [chrM, chr1, chr2, chr3, chr4, chr5, chr6, chr7, chr8, chr9, chr10, chr11, chr12, chr13, chr14, chr15, chr16, chr17, chr18, chr19, chr20, chr21, chr22, chrX, chrY, chr1_gl000191_random, chr1_gl000192_random, chr4_ctg9_hap1, chr4_gl000193_random, chr4_gl000194_random, chr6_apd_hap1, chr6_cox_hap2, chr6_dbb_hap3, chr6_mann_hap4, chr6_mcf_hap5, chr6_qbl_hap6, chr6_ssto_hap7, chr7_gl000195_random, chr8_gl000196_random, chr8_gl000197_random, chr9_gl000198_random, chr9_gl000199_random, chr9_gl000200_random, chr9_gl000201_random, chr11_gl000202_random, chr17_ctg5_hap1, chr17_gl000203_random, chr17_gl000204_random, chr17_gl000205_random, chr17_gl000206_random, chr18_gl000207_random, chr19_gl000208_random, chr19_gl000209_random, chr21_gl000210_random, chrUn_gl000211, chrUn_gl000212, chrUn_gl000213, chrUn_gl000214, chrUn_gl000215, chrUn_gl000216, chrUn_gl000217, chrUn_gl000218, chrUn_gl000219, chrUn_gl000220, chrUn_gl000221, chrUn_gl000222, chrUn_gl000223, chrUn_gl000224, chrUn_gl000225, chrUn_gl000226, chrUn_gl000227, chrUn_gl000228, chrUn_gl000229, chrUn_gl000230, chrUn_gl000231, chrUn_gl000232, chrUn_gl000233, chrUn_gl000234, chrUn_gl000235, chrUn_gl000236, chrUn_gl000237, chrUn_gl000238, chrUn_gl000239, chrUn_gl000240, chrUn_gl000241, chrUn_gl000242, chrUn_gl000243, chrUn_gl000244, chrUn_gl000245, chrUn_gl000246, chrUn_gl000247, chrUn_gl000248, chrUn_gl000249]

[Amitm]

Hi,

Just one addition to what Chris has already said. There is difference in the mito. sequence in the UCSC version as compared to the b37/ 1000G/ Ensembl ver. So if you stick to 1-22 & X and Y only then replacing/ prefixing 'chr' is Ok.

Else take care of the mito. data. And also the alternate/ unplaced contigs. Those are also different in the UCSC ver.

When I analyze WES data, since its (Agilent) not designed to capture mito. anyways, I just choose 1-22, X and Y. Then the data/ sequence of UCSC is interchangeable smoothly with b37/ 1000G

Answer 1

[Chris Miller]

This is the same as your previous problems. "You'll either need to change the dbSNP file or change your data and reference fasta. The former is probably easier - you'll just need to add "chr" when appropriate, change "MT" to "chrM", and convert between the gl contig names

[igor]

There is now a separate dbSNP download section with "corrected" contig names: ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/GATK/

[Mdeng]

This is pretty useful. THX

[umn_bist]

Thanks for your help, Chris. Yes, this has all been little validation errors due to the main issue of not having the original reference.

I did however get a hold of a working reference genome (ucsc.hg19), its corresponding dbSNP and COSMIC vcf but having gone through the formatting process (sorting, indexing, add read group) and finally getting a vcf file with no mutation detection, I think I will resort to the second best option. Do you have any recommendations other than Mutect2 if I am trying to resort to a single tool? FreeBayes/VarScan2/SomaticSniper? GATK has been a very difficult, time consuming (and eye-opening) experience thus far. Thanks again for your help.

EDIT: I find samtools mpileup function much more comfortable to use (but it seems that it is horrible for somatic variant calling).

Answer 2

[Chris Miller]

If you're only going to run one variant caller, Mutect is probably the way to go

[umn_bist]

Does this stand even if I have (impure) tumor samples with no matching normals? I read that MuTect2 is great for pure tumor samples because it picks up low VAF % but for impure ones, it can be too sensitive (high false positives). Does the fact that I have dbSNP and COSMIC vcf ensure that MuTect is good for my use case? Thank you for your help.

[Chris Miller]

No variant caller that I've seen yet is great at low-VAF calling. Impure tumors are more difficult, because the signal is depressed and closer to the noise level from the error rate of the sequencer/prep. If you push too far down, you begin picking those up get a huge number of false positives. My preference is always for some sort of ensemble calling, followed by filtering, but if you're going to use one caller, I still think that Mutect is a reasonable way to go here.