This directory contains the instances from the SeattleSNP data base
used in our paper for computational experiments.

There are 26 instances, namely:

ABCE1
BDKRB2
EDAR
F10
FUK
FUT1
IGF1
IL15RA
IL22
ITGA8
LIPE
MGP
NFKBIA
PLG
PON3
PROC
PROZ
RELA
SELE
SEMA7A
TF
THBD
TNFRSF1A
TYK2
USF1
ZNF202

Each instance, in turn, yields 4 files, with the naming convention:

XXXX (the main instance)
XXXX.full (its conversion in ternary form)
XXXX.col (the "column" instance)
XXXX.row (the "row" instance)

Example:

ABCE1
-----

For example, ABCE1 is the file downloaded from SeattleSNP site, where
the alleles at each SNP are described in textual form, e.g., 
with the letters corresponding to each nucleotides. (See the SeattleSNP
site for explanations on how these files are organized).

ABCE1.full  
----------

The file ABCE1.full contains in matrix form, the same
information as ABCE1. The file is encoded as follows:

<r> <c>
<first row>
...
<last row>

where <r> is the no. of rows of the matrix (i.e., no. of genotypes),
<c> is the no. of columns (i.e., of SNPs sites) and then the rows
of the matrix follow (each row is a string of length <c>). 
Each entry is either 0 (homozygous for allele 0),
1 (homozygous for allele 1), 2 (heterozygous), or 3 (not known)

ABCE1.row
---------

This file is obtained by ABCE1.full by deleting each row containing
an entry "3". The resulting file is a "normal" haplotypoing instance.
Note that some reductions may still be possible, e.g., deleting identical
rows or columns where only "1" or only "0" appear.

ABCE1.col
---------

Same as ABCE1.row, only that the columns of ABCE1.full in which
some entry "3" appeared have been deleted.