Have questions? Visit https://www.reddit.com/r/SNPedia


From SNPedia

A user asks:

I wonder if the C;C ought to be G;G here?

This is a semi-nasty case.


notice this text:

RefSNP Alleles:	A/C/G/T

So all 4 bases have been observed.

SNPedia allows creating all of these pages

which seems to be the best solution.

23andMe reporting wrong build??[edit]

Part of my Promethesease report (done today) is:


115236057 Position

20121021 Rs time

normal, but old version This suggests your data is from build 36. Build 37 is now the standard, and you can re-download your raw data from 23andMe for the updated calls.

My 23andMe raw data says:

"This data file generated by 23andMe at: Wed Sep 26 [xx:xx:xx] 2012 [....] sequence. We are using reference human assembly build 37."

Is 23andMe reporting its build version incorrectly?

--User:Epsilon4 2013-01-4 05:14 (UTC) 05:14, 4 January 2013 (UTC)

There is a reason this is the one which was picked for http://snpedia.blogspot.com/2012/12/snpedias-nightmare-before-christmas.html

Short answer is, I'm 100% sure you're normal. Long answer, the rules for handling 23andMe vs dbSNP orientation for this one don't work.

What follows is my attempt to make sense of this.

What does your 23andMe raw data say?

It probably says G G.

and when you download your raw data, it also says G G

dbSNP tells us this is on the minus strand. But they are using a newer reference GRCh37 37.1/131. But the relative to the 23andMe reference, its no longer the minus strand. I think. Its late. This one frustrates me. Help from others is always welcome. User:Jlick might recall more and have a useful explanation? --- cariaso 05:59, 4 January 2013 (UTC)

This kind of situation is endlessly confusing. It was A/G plus orientation in b36 and now is C/T minus orientation in b37. However, 23andMe raw data always reports in plus orientation, so it would be A or G. G(+) or C(-) is the normal allele. A(+) or T(-) is the pathogenic allele. Looking at the actual ss records and 1000 genomes results, only A/G(+) or C/T(-) is represented, so I'm pretty sure this is really just normal bi-allelic and the A/C/G/T is an artifact of the orientation change. I've seen this happen with other SNPs where the orientation changed. For the current orientation the genotype pages should be (C;C) normal (C;T) heterozygous pathogenic (T;T) homozygous pathogenic. (G;G) normal (A;G) heterozygous pathogenic (A;A) homozygous pathogenic would be alternatives to cover for orientation issues. I don't believe the other combinations are real. --Jlick (talk) 07:32, 4 January 2013 (UTC)

I've edited the genotype pages as per my previous comment and made the magnitudes and descriptions consistent. I've ignored the 'unexpected' genotypes as I don't believe they are real (or if they are they are very rare). --Jlick (talk) 07:59, 4 January 2013 (UTC)

Amazing how confusing this is! I'm not an expert in genomics but I'm not an idiot. It takes a while to realize that much of the confusion comes from the fact that standards -- all kinds of standards related to genes and genomics -- are obviously still evolving. In any event, my raw data says GG; "Lilly Mendel" and "Greg Mendel" also are reported as GG, so it's unlikely this is freakishly rare. Sorry I missed the blog post -- great explanation, thanks! ----User:Epsilon4 2013-01-4 18:31 (UTC) 18:31, 4 January 2013 (UTC)