multipoint runs

David Curtis dcurtis at hgmp.mrc.ac.uk
Thu Jun 15 04:49:00 EST 1995


NECHIPORUK at CSMC.EDU ("Alex Ext. 4983 - Medicine") wrote:


>Very often in publications I see multipoint graphs with 5,6 evev more 
>markers on one single curve.  Even with two alleles for each markers
>I do non think it would be possible to run so many markers at once.  That's
>why I had my question about combining or scaling a couple runs together.

There's no right way to do this. You should draw two lines on the
graph where the runs overlap, leaving out just the right tail of the
first run and the left tail of the second. Choosing either the higher
or lower lod at each point (to produce a single line) is wrong. The
"true" multipoint probably lies somewhere between the two lines, but
not necessarily. The FASTMAP program tries to combine two-points in a
"fair" way, but again it just gives an estimate of what it thinks the
true multipoint might be. 

>Someone suggested to hyplotype two proximal (distal) markers, but we have 
>multiple crossovers between those markers in the pedigree, how would I deal
>with that?

You  can't. And guessing the haplotypes may not be a good idea anyway.
It's rare that they can be deduced unambiguously.

Did you try downcoding alleles? This might let you use all markers at
once.


Dave Curtis (dcurtis at hgmp.mrc.ac.uk)
Institute of Psychiatry, London 
http://www.iop.bpmf.ac.uk/home/depts/psychmed/general/dcurtis/dcurtis.htm




More information about the Gen-link mailing list