README.data

This README.data file describes the files in the data/ directory that were used
as part of the HaploScore publication:
Durand EY, Eriksson N, and McLean CY. Reducing pervasive false positive
identical-by-descent segments detected by large-scale pedigree analysis.
Molecular Biology and Evolution, 2014 (in press).

The data/chr21.23andMe.map.gz file contains the physical map used when detecting
IBD segments in all pairs of individuals. Each line represents a single SNP and
is tab-delimited with four columns:

1. rsid
2. physical position  (NCBI Build GRCh37)
3. allele1  (ordered alphabetically)
4. allele2


The data/chr21.23andMe.segments.{1,2}.gz files contain all 13,307,562
child-other IBD segments identified on chromosome 21 and analyzed for this paper
and the corresponding parent-other IBD segment(s) if any exist. It exists as two
files to get around GitHub's 100 MB file size limit restriction. Each line
represents a single child-other segment and is tab-delimited with three columns:

1. child-other segment
2. parent1-other segment(s)
3. parent2-other segment(s)

The format of each segment is <start_pos>-<end_pos>(<HaploScore>), where
<start_pos> and <end_pos> are the physical positions of the endpoints of the
IBD segment (inclusive) and <HaploScore> is the result of the HaploScore
calculation between the listed individual and the other individual. Multiple
overlapping segments in a single parent, if they exist, are separated by commas.
The first parent listed (P1) is the one used for the parent-other calculation
based on having more SNPs in the overlapping segment (see paper for details).


Examples:

C:10-100(0.5)	P1:10-95(0.6)	P2:

In the above example, the child (C) matches with an external individual from
position 10 through position 100 and the segment has a HaploScore of 0.5. One
parent (P1) has an IBD segment reported with the same external individual; it
matches from position 10 through position 95 with a HaploScore of 0.6. The other
parent has no IBD segment with the same external individual that overlaps
positions 10-100.


C:200-300(0.1)	P1:195-245(0.2),255-300(0.1)	P2:5-205(6.5)

In this example, the child (C) matches with an external individual from
position 200 through position 300 and the segment has a HaploScore of 0.1. Both
parents (P1 and P2) share IBD with the same external individual in this region.
P1 has two IBD segments in this area, one from position 195 to position 245 
and the other from 255-300. P2 has a single IBD segment from position 5 to 
position 205. Parents are displayed in the given order because the segments of 
P1 overlap more of the child-other segment than those of P2 do.