From dnadavenator gmail.com Thu Mar 8 02:45:50 2012
Subject: Re: create Pedigree applying DATA and huge PEDIGREE files
From: David Henderson <dnadavenatorgmail.com>
To: Multiple Recipients of <angenmapanimalgenome.org>
Date: Thu, 08 Mar 2012 02:45:50 -0600
Being one of the authors of GeneticsPed (and the one who wrote the C++ code
that calculates A-inverse), I'm pretty familiar with it's limitations. If
your data can fit in memory in R, then GeneticsPed might work. The A-
inverse matrix does use sparse storage and half stores the matrix in the
C++ part, but when the matrix is passed back into R it is converted to a
full storage matrix. So, you'll need enough RAM to store both the pedigree
and an n x n (where n is the total number
of individuals in the pedigree) matrix of doubles. There used to be a
version that did sorting, but for various reasons it is not in the
current version of the package. Thus, you'll need to sort the pedigree
first. One of the many things I need to fix in the package.
The pedigreemm package sounds like it's pulling its A-inverse matrix from
C/C++ into a sparse matrix and likely has a better chance of working for
you.
I have to admit I'm also a bit confused about what you hope to achieve.
Dave H
On Wed, 2012-03-07 at 02:07 -0600, Rodrigo Gularte Mérida wrote:
> Hello, There is a package available through Bioconductor called GeneticsPed
> that does what you ask for. Here is the website
> http://bioconductor.org/...csPed.html .
> However, I agree with Gregor, if you have problems building your pedigree
> file with Fortran, it is possible that you may have the same problems with
> R. Have you tried using a shell script with tools such as awk, cut, paste,
> and sed to prepare the pedigree ?
>
> best,
>
> ******************************************
> Rodrigo Gularte Mérida
> Unit of Animal Genomics
> GIGA - R -- Université de Liège
> rodrigo.gularteulg.ac.be
>
>
> On 07 Mar 2012, at 08:13, Gregor GORJANC wrote:
>
>> Hi,
>>
>> On 7 March 2012 07:49, Mahmood Honarvar wrote:
>>
>>> Hello Dear All I'm looking for the R-code [or any package] to create
>>> Pedigree applying DATA and huge PEDIGREE files for Mixed Model Analysis,
>>> Including Iterative Queries at determine Generations. I used secateurs
>>> design by karin mayer but its including problems with Huge data files. With
>>> all respects: Honarvar
>>
>>
>> I am not sure what exactly you want to do and what you were able to achieve
>> with Secateurs, but if Fortran program is choking, then any R package will
>> very likely have the same problems!
>>
>> --
>> With regards!
>>
>> ----------------------------------------------------------------------
>> University of Ljubljana Asst. Prof. Gregor Gorjanc, PhD
>> Biotechnical Faculty www: http://sites.google.com/...gorgorjanc
>> Animal Science Department blog: http://ggorjan.blogspot.com
>> Groblje 3 mail: gregor.gorjanc bf.uni-lj.si
>> SI-1230 Domzale fax: +386 (0)1 724 10 04
>> Slovenia, Europe tel: +386 (0)1 320 38 61
>> ----------------------------------------------------------------------
>
|
