Monday January 13, 2003

Version 2.0.1

Executables with compile date after January 10, 2003 (printed at runtime) will handle 
all pedigree structures.

Avalable Platforms:
Linux intel
Sun Sparc Solaris
Mac OSX unix 
Microsoft Windows Conole 
Contact me if you need other platforms.

The loop code has been turned on although not optimized at the moment. Algorithmic
details aside, the performance of vitesse on loops may be better or worse than
other implementations such as FASTLINK. Vitesse uses the loop breaker technique to
handle loops as opposed to cut sets. The number of loop breakers in a pedigree is 
printed to the screen along with the current count as a guide to the feasibility 
of completing the calculation. 

The child vector algorithm has been temporarily disabled since it does not
work with the loop code at the moment.

NOTE: The command line option to print the lod scores for a LINKMAP run has been 
      changed from -M to "-L m" 

------------------------------------------------------------------------------
Also available is PROFILER which can compute probability distributions of 
multilocus genotypes for groups of individuals. PROFILER is built using 
Vitesse and recursive genotype elimination algorithms and can be used with
the same data and pedigree files used in Vitesse.

Given a subset of individiuals in the pedigree, a set of markers called the 
window, and a subset of markers of the window called the slice, PROFILER generates a
rank ordered list of probabilities of each multilocus genotype vector defined
by the slice conditional on the full pedigree and the window. The probabilities
represent the contribution of each genotype vector to the overall likelihood. 
Each genotype vector consists of two haplotypes. So in effect PROFILER is generating
the full haplotype distribution of the set of inidividuals over the slice. 
You can input as many markers as you like in the data and pedigree file but the 
window size is limited by computation time. PROFILER automatically slides the 
window across the framework map.


More documentation will be forth coming but the basics are as follows:

There are two options enabled in this release. 
1. Generate the profile for each individual in the pedigree separately.
2. Generate the profile for a group of user defined individuals. The
   default group is the individual with 1 in the proband column in the 
   pedigree file. To define your own group put a -1 in the proband column
   for each individual in the group.

Command line options:
-D i  (for the individual)
-D U  (for user-defined group)
-s <val> (slice size)
-w <val> (window size)

Examples.
profiler -d mydatafile -p mypedfile -D i -w 3 -s 2
(profiles each individual using sliding window sizes of 3 and sliding
slice sizes of 2 within the window)

profiler -d mydatafile -p mypedfile -D U -w 2 -s 2
(profiles the group using sliding window and slice sizes of 2 )

Output is to the screen and also the file "profiler_results".
Genotype vectors use ordered genotypes A | B representng a 
paternal A allele and maternal B allele. Thus, multilocus 
genotypes represent paternal and maternal haplotypes.

Applications of PROFILER to perform haplotyping, pedigree error analysis,
approximate the likelihood and IBD sharing statistics are under development.

PROFILER does NOT work with looped pedigrees at the moment.


The reference for PROFILER is
 O'Connell JR (2000) PROFILER: a program to compute cumulative probability
 profiles with applications to pedigree analysis. Amer J Hum Gen 67:A328
______________________________________________________________________

NOTE new contact e-mail:
joconnel@medicine.umaryland.edu