Somatic Mutations in dbSNP

[qqcandy]

We are looking for somatic mutations in tumor by exon sequencing with paired normal samples. Our original plan is to sequence 20 tumor samples and filter out polymorphisms/mutations in dbSNP129. Then among the remaining polymorphisms/mutations that are non-synonymous, we are going to choose those observed in at least 3 out of 20 samples to validate by PCR, as well as to see if they are somatic or germline by PCR the paired normal. But since the dbSNP has somatic mutation, what is the chance of mutations related to cancer progression (something we are really looking for) be filtered if we filter out dbSNP mutation? Any thought or suggested are appreciated ~~

The alternative is to sequence the normal paired sample exactly the same way as the tumor, but it is probably much more expensive...

[qqcandy]

http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=helpsnpfaq&part=Submit.Submitting_Various_D

"""
Q: Can we submit SNPs that might be somatic mutations? Problem is, we don’t know which are germline SNPs and which are the mutations since we didn't sequence matched normal DNA.
A: dbSNP accepts SNPs and mutations. Make sure that you state in your methods, however, that you have no way of knowing which SNPs are somatic and which are germline SNPs/mutations.
"""
To think about it, for the sample we consider normal, "normal" just means that the tissue derive the DNA from doesn't look abnormal, it doesn't mean it doesn't harbor somatic mutation... This means we can define somatic confidently, but we never know whether the germline is true germline (i.e. come from the parents), unless we sequence the parents ? !

[krobison]

dbSNP does have annotation as to what sort of study the SNP was found in -- you could just filter out those SNPs which come from large studies (HapMap, 1000 genomes, etc)

[qqcandy]

dbSNP does have annotation as to what sort of study the SNP was found in -- you could just filter out those SNPs which come from large studies (HapMap, 1000 genomes, etc)

Thank you for the suggestion. Could you explain why filter out SNP from large studies? They are more likely to include somatic mutations, or they have little annotation? Thanks

[krobison]

These large studies are looking at germline polymorphisms which are common in the human population. For many of these there should be both frequency statistics for the alleles & information as to the sample size. So, if you are interested in somatic mutations anything that is in dbSNP and was found by one of these large studies is almost certainly an inherited polymorphism.

E.g., ERBB2 P1170A, which is a well studied polymorphism in an known oncogene

http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=1058808

(obviously, I'm cheating by using NCBI's nice GUI; doing this programmatically is clearly what you need)

[spenthil]

Sounds like what the Shendure lab did last month:

http://www.nature.com/ng/journal/v42/n1/abs/ng.499.html

[qqcandy]

Sounds like what the Shendure lab did last month:
http://www.nature.com/ng/journal/v42/n1/abs/ng.499.html

Yes, except they knew they probably had only one mutation to look for and that mutation is probably in all disease samples; while in our case, there are for certain more than one somatic mutations in cancer pathogenesis and quite likely none of them will be in all cancer samples...

BTW: They filtered out dbSNP129 and it may be a problem:

"""
But in another cohort (4 individuals from 3 kindreds with Miller syndrome, a rare developmental disorder) Dr. Shendure and colleagues discovered the danger of overfiltering. They removed all variants from dbSNP 129, but when they limited the scope to only mutations predicted to be “damaging” or “deleterious”, the number of genes dropped to zero. Apparently the deleteriousness of at least one of the causal mutations wasn’t predicated correctly.

Obviously, the need is for better filters of common variants. But with projects like the 1,000 Genomes in full swing, I wonder, will filtering out using dbSNP get better, or worse? Already, as Shendure pointed out, certain genes have basically a SNP reported at every position. I know that TP53 does. What’s more, with the advent of next-generation sequencing, I hate to tell you, but people are going to be reporting a lot of false positives. I guarantee it. So when you filter all of the variants, you might actually remove the ones you’re looking for.
"""

http://www.massgenomics.org/2010/01/capture-and-subassembly-with-jay-shendure.html

[NGSfan]

what about taking the dbSNP, comparing it to COSMIC mutations and removing the entries in dbSNP that overlap with COSMIC?

At least you would not filter out known somatic mutations.

[hanifk]

I also dont have matched normal sample sequencing, I also want to minus dbsnp, I also worried that somatic mutations are filtered.

[Richard Finney]

Any progress on this front?
Any better annotation in dbSNP?
Any better methods?

[AStretton]

If you only want to remove common SNPs I would recommend that you use something like HapMap or 1000genomes not dbSNP, and also consider your study population as what maybe a very common SNP in one ethnicity may not be in another. dbSNP has all sorts of mutations uploaded into it and not all of these would be considred common variations.

[LiLin]

You may filter with dbsnp137, flag "SAO",maybe helpful.

[zjrouc]

how did you use dbsnp

Hi,

Would you please tell me in detail how did you filter out polymorphism using dbsnp ? i am newbie and learning this.

Thanks.

[m_two]

Consider using Exac to identify germline SNPs in coding regions.

The Jan 2015 release consolidates exomes data from 60,706 individuals.

http://exac.broadinstitute.org/

ftp://ftp.broadinstitute.org/pub/ExA...se/release0.3/

[m_two]

dbSNP data is available in VCF format on FTP and pretty well documented but not as extensive or up-to-date as ExAc. It never hurts to have a backup.

http://www.ncbi.nlm.nih.gov/variation/docs/human_variation_vcf/

ftp://ftp.ncbi.nlm.nih.gov/snp/organ...uman_9606/VCF/

You may want to parse the 1000 Genomes CAF info from the file to identify germline events. Be aware of the caveats described in the documentation.