Phasing Data With Beagle

Question

[Tim]

Hey

I need phased genotype data for another statistic I want to calculate and I decided to phase my data with BEAGLE. Before even starting to phase I extracted with PLINK the loci which have a genotype rate of 98% or above sot hat there are only a few loci which woudl be candidates for imputation. My overall set is now about 120k markers big (SNPs). I wanted to know whether I can just phase this data set or hether youw ould recommend keeping the whole data set (all markers) but accept a higher imputation rate? Imutation can cause all sorts of biases - is 120k markers enough to infer haplotypes?

Tim

[aa.abied89]

why i need to phase my data by Beagle to used selscan software ?

[aa.abied89]

i need tutorial or i think manual its better for selsacan and beagle software?

[finswimmer]

Hello aa.abied89 and welcome to biostars,

Answers should ONLY be used to respond to the original question at the top of this page and providing (full) solutions. Use the ADD COMMENT or ADD REPLY buttons above to respond to someone, ask for clarifications, request more details etc, This is why I moved your posts to a comment.

But I think, if you have specific questions to beagle, you should start your own thread if these question have nothing to do with the one in the OP.

fin swimmer

Answer 1

[zx8754]

See Avsmith's answer regarding pre-imputation QC on this thread How to analyze imputed GWAS data

If you wish to be strict on QC, you can follow the tutorial on this paper: http://www.ncbi.nlm.nih.gov/pubmed/21085122

Also, this paper suggests that maybe it is not necessary to QC before imputation... http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3083623/

Regarding the question, if 120K SNPs are enough to infer haplotypes: from recent experience, we imputed 1.2Mb region with ~500 typed SNPs. At INFO>0.5 we have now about 3000 imputed SNPs with 98.9% concordance.