In article 16180 at news.unige.ch, mike at medsun.unige.ch (Mike Morris) writes:
> Not a common complaint, I guess. But I cannot understand (or believe)
>a result we recently had from LINKAGE 5.1.
>> We have a consanguineous family, with a monogenic AR disease (locus
>1). We have two perfectly linked markers (2 and 3), which complement
>each other in terms of informativity. The markers are at most 1% apart
>(no rec. in this family).
>> I ran MLINK, "Lod score table", with the following results:
> Locus order Rec fractions Rec varied lod max (always theta=0)
> 1 2 .0 1 3.1
> 1 3 .0 1 1.3
> >>> 1 2 3 .0 .01 1 16.1 <<<
>> As I understand this, either I have screwed up somewhere, or the
>program interprets the third line as "1 vs. superlocus 2+3", and I
>get a genuine result. But 16 seems much too high!
>> Can anyone clarify this for me?
>> Mike Morris
>>*******************************************************************
>Division of Medical Genetics tel (Switzerland) (22) 702.56.94
>CMU, University of Geneva fax (Switzerland) (22) 702.57.06
>Geneva, Switzerland email mike at medsun.unige.ch>>
Its not possible to say what your lod score should be without knowing
your pedigree structure, of course.
Why don't you simulate the maximum lod score that you could get
by inputting a single hypothetical 'magic marker' which is fully
informative (from what you say, this would be equivalent to
numbering each of the haplotypes of your existing data as a single
locus, reasonable if there is no recombination with such close markers).
The marker allele frequencies could be very important - try a wide range
for the disease-linked allele.
This is not a tidy theoretical answer, but should give you the information
you need.
