Faster linkage analysis with FASTLINK

wijsman at max.u.washington.edu wijsman at max.u.washington.edu
Mon Aug 30 17:13:35 EST 1993


We have finally gotten Bob Cottingham's new improved version of LINKMAP
(FASTLINK) up and running on our computers, and have performed a test run on
one of our Alzheimer's disease pedigrees.  So far we have only built &
tested the "slower" of the two versions of the program (and we haven't
tried any of the other programs in the package).  We need to do some
systems revisions in order to obtain the memory necessary to try the
fastest version.  It is not clear when we will have time for that, but it
won't be happening immediately.

The test run we performed took approximately 1/16 of the time necessary to
do the same analysis with v. 5.1 of the LINKAGE package.  For the problems
we are working on, this is a significant increase in speed (and thus
feasibility) - see below. 

The pedigree:  9 individuals on whom we have marker data out of 22
individuals in the family; 8 individuals are affected (we have marker data
on 2 of these with marker genotypes inferrable in a third affected person
on the basis of spouse and kids).

Loci in the analysis:  3 markers, with 4, 3 and 6 alleles respectively,
after downcoding to a minimum necessary number of alleles, plus the disease
with 2 alleles and 6 liability classes (144 total possible haplotypes). 

Time needed to compute the likelihoods for ONE recombination fraction on a
Vaxstation 4000:  approx. 8 CPU hours with the standard LINKMAP program; 
approx. 36 CPU minutes with the FASTLINK program.  Thus a full analysis
of this one pedigree across the region spanned by these markers changes
from a 1-week run to an overnight run!

Ellen Wijsman
Div of Medical Genetics, RG-25
and Dept of Biostatistics
University of Washington
Seattle, WA   98195
wijsman at u.washington.edu



More information about the Gen-link mailing list