|
Post by Eleonora Porcu on Dec 13, 2013 14:38:43 GMT
Hi,
in the point 5 of the website's FAQ, you wrote:
If you are interested in the variance explained by a subset of SNPs in family data, you could fit the genetic relationship matrix (GRM) estimated from these SNPs along with a matrix of pedigree structure using the option --mgrm when running the REML analysis (--reml).
How can I create the matrix of pedigree structure with GCTA? which format should it have?
Thanks Eleonora
|
|
|
Post by Jian Yang on Dec 14, 2013 7:35:25 GMT
|
|
|
Post by JA on Feb 18, 2015 23:03:50 GMT
Thanks for the useful article, I'm interested in estimate the variance explained by SNP in family data. Number of genotyped SNPs is very small and the only matrix I can use is based on the pedigree structure. Can you please provide an example of the matrix that I should used instead of the GRM based on SNP?
Thanks
|
|
|
Post by Jian Yang on Feb 25, 2015 14:58:03 GMT
You could edit your a GRM using the format below.
test.grm.gz no header line; columns are indices of pairs of individuals (row numbers of the test.grm.id), number of non-missing SNPs (you don't have that in your case so you can use any positive number, e.g. 10) and the expected of genetic relatedness (e.g. 0.5 for full sibs).
1 1 10 1 2 1 10 0.5 2 2 10 1 3 1 10 0.5 …… test.grm.id (no header line; columns are family ID and individual ID)
011 0101 012 0102 013 0103 ……
Please be aware of that in this case you just use GCTA as a REML engine. The estimate from this analysis is equivalent to that from a pedigree based analysis rather than the variance explained by SNPs.
|
|
|
Post by JA on Mar 4, 2015 11:56:01 GMT
Thanks so much for your reply, So basically the estimate from this analysis will equal the heritability of the trati !! Can you please suggest any other method that can be used to estimate the variance explained by the genotyped SNPs (number of SNPs is 43), taking into account that I can only use the pedigree to produce the GRM..
|
|