Do variant callers miss rare variants in reference? [Super Explanatory Image Inside]

[Nikvailo]

Generally variant calling programs (such as GATK-UnifiedGenotyper) look for differences between reference genome and submitted sequence. However, we all know that reference genome consist rare variants in various positions. If submitted sequence has this rare form of that variant, variant call doesn't see it and doesn't report at all. However, that variant could be very precious for analysis.

So, how can I solve this problem? What should I do to not miss that rare variants even they are in reference genome?

[ronton]

This is why large samples are important in genomics. 1000 genomes for example. Each person's genome has variability compared to the next, not necessarily pathogenic or disease-related variants, just variation. So, looking at variants in the context of per 100 or 1000 individuals may be more useful than comparing one individual to one reference genome.

Hopefully hg19 is annotated with rare variants and/or alternate alleles, rather than one single genome sequence.

You actually supply 'known information' from several databases including 1000G to GATK, but I am not 100% sure how this factors into variant calling.

[Bukowski]

http://ashleylab.stanford.edu/tools/synthetic_reference.html

[dpryan]

Cross-posted on the biology stack-exchange, where I already replied. Please don't cross-post!

[Nikvailo]

@ronton
Thank you so much.I'll look for alternate alleles of hg19.

@Bukowski
That could be exactly what I look for. Thank you so much.

@ dpryan
Thanks for the answer. But Why shouldn't I? I want to reach people as many as possible. Actually I asked the same question on Quora. I don't think that it wrong. Seqanswers and Stackexchange are different platforms and may have different users. So, a user that is member of only single platform may miss my question. Cross-posting increase the possibility to get answer. (P.S This is also cross-posting answer.)

[dpryan]

I replied on the biology stack exchange. The short answer is that it wastes community resources.

[Nikvailo]

Actually, I'll share the answer that I found in every platform that I asked question. So, everyone in both platform will learn the answer. More people will benefit. It's better right?

[pmiguel]

Actually, I'll share the answer that I found in every platform that I asked question. So, everyone in both platform will learn the answer. More people will benefit. It's better right?

I completely agree with you Nikvailo. I can't imagine any harm in such "cross-posts".

--
Phillip

[dariober]

I completely agree with you Nikvailo. I can't imagine any harm in such "cross-posts".

--
Phillip

(The thread is degenerating)
IMHO cross posting is fine for the reasons Nikvailo mentioned, especially if the question sits in between different fields. However, if you cross post you should explicitly warn the reader by adding links to where the same question was asked. Then as usual, good common sense applies...