From mari0409_at_ufl.edu Wed Feb 15 13:42:08 2006
Subject: discrepancy between two-pt and multipoint analysis
Date: Wed, 15 Feb 2006 13:49:16 -0500
From: "Mariasegaram,Maxy" <mari0409_at_ufl.edu>
To: Multiple Recipients of <angenmap_at_animalgenome.org>
Dear Colleagues,
I am involved in mapping the gene responsible for the slick hair coat in
Senepol derived cattle. The gene follows an autosomal dominant mode of
inheritance, and the phenotype is being analysed as a binary trait in
Superlink. My questions concerns the localisation of the gene using two-
point and multi point linkage analyses. There appears to be a 5 cM
discrepancy between the two methods in terms of the predicted location. I
have checked my genotype data with Crimap's chrompic option to identify
potential double recombinants and pedigree errors, but have found none. Has
anyone else encountered similar problems and how do you resolve it?
Thanks in advance for any suggestions.
Cheers
Max
Maxy Mariasegaram
USDA/ARS
Subtropical Agricultural Research Station
22271 Chinsegut Hill Road
Brooksville
FL 34601
Tel: + 1 352-796-3385 x233
Fax: + 1 352-796-3520
|
From jheox_at_ntlworld.com Mon Feb 20 14:16:24 2006
From: jheox_at_ntlworld.com
Subject: Re: discrepancy between two-pt and multipoint analysis
Date: Mon, 20 Feb 2006 19:07:28 +0000
To: Multiple Recipients of <angenmap_at_animalgenome.org>
Could I make some points that are not, I think, in doubt.
1. Two-point linkage analysis is based on well established assumptions and
will only give problems if segregation is disturbed - most package programs
assume Mendel's first law obeyed and proceed to estimate deviations from his
second law.
This is simple to check with dialelic loci. The statistical problem is
similar to that presented by the human ABO blood-group locus.
2. Multi-locus analysis assumes the order known, as well as making
assumptions on absence of sex difference in recombination and, in some
programs, the 'backward look' of haplotyping parents from children.
If it works then it should produce similar answers with less reliability on
the mean but a lower variance, in which case it could have substantial
advantages once checked against methods based on sounder assumptions and,
unless marker order based on physical mapping, sounder data.
Disturbed segregation can be due to various causes, of which the most
obvious is the presence of recessive lethals from deficiencies. The
established testing procedures do not distinguish homozygotes from
hemizygotes. In some human disorders haplo-insufficiency can mimic dominant
inheritance.
As I am interested in mapping deficiencies in mammals I would welcome
hearing more.
John Edwards
|
Go back to the