Re: BimBam
Yongtao Guan
You have to use panel in order to do imputation.
The data for hapmap panel can be downloaded from the bimbam resource page.
When you use "-i 1" it means "do imputation to get mean genotypes,
then calc BF based on the mean genotypes".
For "pval" option, it has to follow a number that is larger than 100.
(it indicates doing how many permutations to get p-values for each
snp). Note for a large chromosome the pval option can be slow. I would
recommend you only do this for SNPs of large BFs.
The following command should work.
./bimbam_linux_64b -g ch13.geno.txt -p y.pheno.txt -o ch13.out -sort
-g panel.chr13 -p 0 -i 1 -pval 1000
To get a imputed SNPs, you may use -wmg or -wgd, the former is to
write mean genotype, the later is to write genotype distributions.
Again, you have to include panel when use these options. These
commands are indeed documented in the manuel:
http://quartus.uchicago.edu/~yguan/bimbam/instruction/node31.html
Last, you might want to try the new version of the bimbam, which can
be downloaded directly from here:
http://quartus.uchicago.edu/~yguan/bimbam/bimbam_lin
Grant
On Sun, Oct 19, 2008 at 11:09 PM, <Huwaid...@csiro.au> wrote:
> Hi Dr.Guan,
>
> I tried to use BimBam on my data on one chromosome at a time (as
> recommended in the manual ) with the following code
>
> "./bimbam_linux_64b -g ch13.geno.txt -p y.pheno.txt -o ch13.out -sort "
> it worked fine. but when using -i 1 -pval 5, for example
>
> "./bimbam_linux_64b -g ch13.geno.txt -p y.pheno.txt -o ch13.out -sort -i 1
> -pval 5 "
> I got:
>
> no input prior or priors not matching, use default priors
> Warning: can't open rs-pos file, proceed with care.
> ## num genotype files = 1
> ## num phenotype files = 1
> ## num of diploids = 200
> ## num of haploids = 0
> ## num SNPs = 36590
> ## bad SNPs = 582
> ## num Phenotype = 1
> ## num Cohort = 200
> ## num Panel = 0
> nSubPop = 1
> Calc BFs Based On Mean Genotype:
> [>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 100%
> The following files have been written in your output subdirectory:
> ch13.out.snpdata.txt
> ch13.out.single.txt
> ch13.out.summary.txt
> ch13.out.log.txt
> random seed = 1224471880
>
> [1] Exit 1 ./bimbam_linux_64b -g ch13.geno.txt -p
> y.pheno.txt -o ch13.out -sort -i 1 -pval ...
>
>
>
> the log file:
> ##
> ## COMMAND: ./bimbam_linux_64b -g ch13.geno.txt -p y.pheno.txt -o ch13.out
> -sort -i 1 -pval 5
> ## num genotype files = 1
> ## num phenotype files = 1
> ## num of diploids = 200
> ## num of haploids = 0
> ## num SNPs = 36590
> ## bad SNPs = 582
> ## Write single SNP Bayes `factors (mean genotype) into a file.
> ## total imputation = 1
>
> my other question is: can we get the imputed genotype data? I mean imputing
> the missing genotypes in my data.
> I know that fastPHASE does that but Matthew Stephens recommended using
> BimBam. I couldn't find that in BimBam manual.
>
> would you please advise on both issues?
>
> Thank you.
> Huwaida Rabie, Ph.D.
>
--
Yongtao Guan, PhD
Postdoctoral Fellow
Dept of Human Genetics
University of Chicago
http://home.uchicago.edu/~ytguan/