Delivered-To: crimap-usersanimalgenome.org 
From: "Vincent Prieur" <vincent.prieurlimousine.org> 
To: Multiple Recipients of Cri-Map Users <crimap-usersanimalgenome.org> 
Subject: RE: Marker selection 
Date: Tue, 04 Oct 2016 10:18:08 -0500 

Hi,

I actually use CRI MAP to estimate some genetics maps using thousands of SNP
markers easily.

First, you may want to set your .gen file by separating the families.  This
will help to make CRI MAP to run a lot faster.

Then I estimate a genetics map for each chromosome. It is very fast if I only
have parents and product genotyped (150 animals spread in 40 families, it
takes a few seconds for the biggest chromosomes, the 1) . When you have some
trios and bigger families, it can be a little bit longer. Maybe you would
separate your chromosomes in two or more subset overlapping each other.

Then once you get the genetic position, you can construct a loess model by
fitting your genetic map on the physical map.

And the last step, applying a prediction to the full data set will give you
the genetics map for every markers. (I did it to build genetics map on the
54K bead chip on sheep and cattle)

My article is not yet published on this. But if you use this method, I will
sent you the reference of my thesis that you could use for the citation please.

Many thanks.

 
Vincent Prieur 
Ingnieur R&D  
Service Technique 
France Limousin Slection - Ple de Lanaud  87220 BOISSEUIL 
http://www.limousine.org  

 
-----Message d'origine----- 
De: Littrell, John [mailto:jlittrellmcw.edu]  
Envoy: mardi 4 octobre 2016 16:02 
: Multiple Recipients of Cri-Map Users <crimap-usersanimalgenome.org> 
Objet: Marker selection 

Hi all, 

Thank you for your help with my last questions. This time I have a question
about marker selection. We have more markers than we can fine map - 10,000 on
the smallest chromosome with only 600+ individuals in 60+ nuclear families. I
figured we would start with a reasonable number of markers and continue to
use more until build could no longer add more markers to the map (at the same 
LOD) because there were not enough meioses. Well, we are now at 4000 markers 
and the map continues to grow. It takes nearly two months to run. I need a 
better strategy for the larger chromosomes. 

I am looking for a way to select the most useful markers instead of a lot of
markers. Here are my questions: It is my understanding that using nuclear
families means that we have no phase information and as a result the order of
marker informativeness created by prepare is not useful. Is that correct?
There are several families in the pedigree that have grandparents 
(unfortunately, usually for just one of the parents). Should we use those to 
get an idea of informativeness or would that bias marker selection to those 
families? Or would using MAF as an indicator of informativeness be a good 
idea and the range of 0.25-0.50 reasonable? 

Thank you for your help, again.

Jack 
Genetic map enthusiast