-----------------------------------------
Haplotype inference in half-sib families
-----------------------------------------
by Aurelie Favier, INRA, Toulouse, France
-----------------------------------------

Input file format
-----------------

toulbar2 Haplotype Inference input files are the pedigree file and the map file. The pedigree file (*.pre) must contain the following information for each individual:

 * a pedigree (family) number
 * an individual identification number, or id
 * father's id number
 * mother's id number
 * sex 
 * first allele  number
 * second allele number
 * ...

warning! only one family is allowed per file in the current toulbar2 version. if you have several families, then you have to create one pedigree file for each family.

Father (must be the first individual in the file) and mother id's are 0 for founders, or other members of the pedigree for whom information on parents is absent. Partial information is allowed (when only one parent is known). The sex field is coded 1 for males and 2 for females. The genotyping data consists of two unordered allele numbers (positive integers). These are codominant alleles at a single locus. An unknown genotype is coded as 0 0.

The map file (*.map) must contain the genetic position in Morgan for all the markers present in the pedigree file


Example with 1 family, 4 individuals (1 sire and 3 offspring), and 5 markers
----------------------------------------------------------------------------

pedigree.pre:

 1  1  0  0  1      1 2 2 2 1 1 2 2 1 1
 1  2  1  0  1      2 1 2 2 1 1 2 1 1 2
 1  3  1  0  1      1 1 2 2 1 1 1 2 1 2
 1  4  1  0  1      2 2 2 2 1 1 2 1 1 1

position.map:

 0.000 0.025 0.050 0.075 0.100

Try the following command :

 toulbar2 pedigree.pre positions.map H w

The software output will be:

 4 informative individuals found (either genotyped or having a genotyped descendant).
 Read 5 variables, with 2 values at most, and 4 constraints.
 Reading and creating problem time :	 0
 0 unassigned variables, 0 values in all current domains and 0 constraints.
 New solution: 0 log10like: -0.966877 logProb: -2.22632 (0 backtracks, 0 nodes, depth 0)

 Optimum: 0 log10like: -0.966877 logProb: -2.22632 in 0 backtracks and 0 nodes, and 0 seconds.
 end.

and look at the output file haplotypes.

 sire 1 
 1  1  2  2  1		      ( haplotype 1 )
 2  1  1  1  2		      ( haplotype 2 )
 log10like	: -0.966877   ( the log-likelihood in base 10 )
 logProb	: -2.22632    ( the log-likelihood )

For large pedigrees with a lot of markers (e.g. ChrX? instances, see below), use the BTD search method instead of DFBB used by default in toulbar2. For that, add option B1:

 toulbar2_09 pedigree.pre position.map Hw B1

Use the v option to see the non-zero value of the matrix W. Try the following command :

 toulbar2_09 pedigree.pre position.map Hw v

The software output will be:

 ../../../Toulbar/toulbar2_09.01/bin/Linux/toulbar2  version : 0.9.1.1, copyright (c) INRA 2010
  4 informative individuals found (either genotyped or having a genotyped descendant).
 sire is individual 1
 sparse matrix : 
 W0,2	 =	 -0.749141
 W2,3	 =	  0.922272
 W2,4	 =	  0.749141
 W3,4	 =	  0.922272
 constant K   =	 -8.91197 
 Read 5 variables, with 2 values at most, and 4 constraints.
 pedigree ub: 76971419
 Reading and creating problem time :	 0
 Variable elimination during search of degree <= 3
 0 unassigned variables, 0 values in all current domains and 0 constraints.
 New solution: 0 log10like: -0.966877 logProb: -2.22632 (0 backtracks, 0 nodes, depth 0)

 Optimum: 0 log10like: -0.966877 logProb: -2.22632 in 0 backtracks and 0 nodes, and 0 seconds.
 end.


Datasets
--------

pattern for each directory : numberOfDescendants_numberOfMarkers

Human chromosome X
http://carlit.toulouse.inra.fr/homes/afavier/Human_chrX.tar.gz

For comparison between exact methods
http://carlit.toulouse.inra.fr/homes/afavier/exact_methods.tar.gz

For comparison between approximate methods
http://carlit.toulouse.inra.fr/homes/afavier/approx_methods.tar.gz

For comparison with and without linkage desequilibrium
http://carlit.toulouse.inra.fr/homes/afavier/DL_vs_noDL.tar.gz


-------------------------------------------------------------------------------------------------
Aurelie Favier, INRA, Toulouse, France @ 2010
aurelie.favier@toulouse.inra.fr
