This directory contains genotypes inferred using the method
described in

Burdick JT, Chen WM, Abecasis GR and Cheung VG (2006). In silico
method for inferring genotypes in pedigrees. Nat Genet 38:1002-4.

http://genomics.med.upenn.edu/genotypeinference/Nat.Genet.vol.38-pp.1002.pdf


To do the inference, we started with HapMap genotypes for 60 CEU individuals
(the ten families which had HapMap genotypes for four grandparents and
two parents), using HapMap build 16c1 and build 21.  We omitted
chromosomes X and Y.

Next, we computed IBD sharing in those ten families, using genotypes from
The SNP Consortium (TSC) for 1,978 markers, combined with a set of genotypes
for 4,586 markers generated using the Illumina platform.

Then, we looked for cases in which we knew which grandparent transmitted
an allele, and we can determine which allele was transmitted.  For instance,
if the IBD information indicates that the maternal and paternal grandmothers
had each transmitted an allele to a child, and we can tell from the HapMap
genotypes that "C" and "T" had been transmitted, then the child's genotype
must be "CT".  (Of course, if a parent's genotype was "CC", then we know
that parent transmitted a "C" to the child, even if we lack IBD information
at that location.)

The resulting genotypes are for 78 CEU children in rel 16c1 and rel21, or 73 
CEU children in rel22 in those ten families (see corresponding samples lists).  
Missing alleles are represented by "N".  Note that in some
cases, the inference program could only infer one allele.  So, for
instance, genotypes of "NT" or "TN" mean that only one allele ("T")
could be determined.  (This could be because of missing IBD information,
or because we couldn't determine which allele was transmitted.)  "NN"
represents cases in which the inference program couldn't determine
either allele.

The program used to infer genotypes is available at

http://genomics.med.upenn.edu/genotypeinference/


Josh Burdick
jburdick@gradient.cis.upenn.edu
20061009

==

The genotype files contain the following columns:

snp_rsid: The SNP rsid 
alleles: The alleles according to dbSNP
chrom: chromosome
pos: position on the chromosome
strand: strand releative to the genome
build: genome build
source_center: The center which provided the original genotypes in HapMap
source_assayLSID: The assayLSID of the original genotypes in HapMap
protLSID: protocol lsid for these inferred genotypes

followed by the genotypes for all the inferred individuals.

== 

The file 'sample_list.txt' contains information on the individuals in the list.

There are the following columns:

sample_name: Sample name (nothe this is the cel line identifier.  For
	     comparison to HapMap ids, substitute 'NA' for 'GM'
ceph_family: CEPH family number 
ceph_individual: individual number in CEPH family
gender: sex of individual
relation: familial relationship within the pedigree
inferred: 'yes' for those with inferred genotypes.  'no' marks the
	  family members typed within the HapMap project.