From jillian.maddox alumni.unimelb.edu.au Thu Apr 4 18:28:58 2013
From: Jill Maddox <jillian.maddoxalumni.unimelb.edu.au>
To: Multiple Recipients of <crimap-usersanimalgenome.org>
Subject: RE: General questions of using crimap
Date: Thu, 04 Apr 2013 18:28:58 -0500
Hi Gang and other CRI-MAP users
There was a parallel version of CRI-MAP that used PVM (crimap-pvm) but this
method is no longer widely used and that version of CRI-MAP hasn't had all
the memory fixes etc that the current version has so won't run with large
data sets. We had planned to parallelize CRI-MAP (OpenMPI) after we have
the new map function working and fixing up some other bits as this was a
more efficient use of our resources than doing things the other way round.
However, the fixing of other bits has turned out to be more complex than
expected and is taking longer. In the meantime the funding for this work
has run out so the parallelisation is looking less likely unless someone
wants to volunteer to do it or to provide some funds. Please let us know if
anyone wants to do the parallelisation (we would need to coordinate which
version of code gets used for the changes as the source code we are working
on has a number of differences to the versions that are publicly available)
or provide some funds for us to do it.
I believe that you (Gang) said that you had only 300 animals in your
pedigree and you have 50,000 SNPs. This number of informative meioses is
not a sufficient number to build a high density linkage map de novo
especially given you are using SNPs. If you already have a known marker
order then you could check the order with CRI-MAP or other linkage mapping
programs or you could investigate using phasing software (need markers to
be ordered for this).
Best wishes from
Jill
At 01:49 AM 5/04/2013, Li, Gang wrote:
>
>Hi Jill:
> Thank you very much. This problem have been resolved.
> Because my data is a big matrix, more than 50,000 markers. It is still
>take long time to run twopoint calculation. I notice only one
>thread is using
>for crimap running. Do you have any suggestion to run crimap
>applying multiple
>thread?
> I appreciate your help.
>Best regards,
>Gang
>
>
>
>
>-----Original Message-----
>.From: Jill Maddox [mailto:jillian.maddoxalumni.unimelb.edu.au]
>.Sent: Wednesday, April 03, 2013 6:55 PM
>.To: Li, Gang; Multiple Recipients of
>.Subject: Re: General questions of using crimap
>
>Hi Gang
>
>There are 2 ways to run prepare without having to do keyboard input.
>
>The first is to just use lispcri rather than crimap to run prepare. If you
>don't specify otherwise lispcri is built automatically when CRI-MAP is made
>- just make sure it is in your path. If it hasn't been built then do make
>lispcri in the source code directory. e.g. lispcri n prepare > chrn.prepare
>where n is the number of your chromosome Make sure that there is not
>already a chrn.dat file in the directory as if there is then lispcri will
>hang.
>
>The second way is to create an answer file for crimap with the answers that
>you want to use. e.g. input file called ans containing the following lines
>N N N N 8 Y
>
>replace the 8 with whichever CRI-MAP option you want to do then invoke with
>crimap n prepare chrn.prepare Note that the same caveat re pre-existing dat
>files applies. If there is a pre-existing dat file either delete it or set
>up a slightly different answer file to allow for it so that the answers in
>the file match the questions asked.
>
>Regards
>
>
>Jill
>At 02:17 PM 3/04/2013, Li, Gang wrote:
>
> >Hi Every One:
> > I am a new user of crimap.
> > The data I have are 50,000 SNPs from a snp array
> >provided by more than 300 individuals, it is a big matrix. Family
> >structure are kind of complicate. Individual samples tested using this
> >SNP array come from multiple families with complex structure. Original
> >parents, parents and offspring are mixed together.
> > 1. Problems confused me now is the method to handle
> >this data. After checking the literature, crimap is an appropriate
> >software to build the linkage map using my data. But if I only used a
> >partial data (with high informative markers), it is still very very
> >slow to run everay thing, even twopoint.
> > 2. I try to separate the data to multiple subsets for
> >each chromosome and use a sliding windows method to do a quick check of
> >the published old maps. I can not run crimap in a batch model using a
> >perl script, because of option 'prepare' requires several interrupt to
> >ask users' options.
> >Is there any way I can ignore that or let crimap automatically pick up
> >options?
> >Thank you so much for your any helps. Gang
> >~~>bioinfo-teamanimalgenome.org>Program - Supported by
> USDA/NIFA>crimap-users-requestanimalgenome.org
>
>
>***************************************************************
>
>Jill Maddox 16 Park Square Port Melbourne, 3207 Australia phone: 03 9646
>0428 E-mail: jillian.maddoxalumni.unimelb.edu.au
>
>***************************************************************
>
>
***************************************************************
Jill Maddox 16 Park Square Port Melbourne, 3207 Australia phone: 03 9646
0428 E-mail: jillian.maddoxalumni.unimelb.edu.au
***************************************************************
|
