==File contents and formats==

Each directory (001, 002, ..., 010) contains the following data for one
individual:

* GM#####.bed.gz - Affymetrix 500K scan results, in BED format 

* pgp#.cns.fq.txt.gz - Exome sequencing results, in FASTQ format, placed
  against the 6.7Mbp target reference library 

* 55kpgp#.bed.gz - Exome sequencing results for PGP-10 SNP positions only,
  in BED format placed against the 6.7Mbp target reference library 

* hg18pgp#.bed.gz - Exome sequencing results for PGP-10 SNP positions
  only, in BED format placed against the HG18 reference build 

===Notes on BED format used===

* BED coordinates are relative to the HG18 reference assembly.   


* The "name" field in the Affymetrix scan results is given as X>YZ:D:
  (Example: "A>AG:rs1565744:").

X - the "major allele" specified in the Affymetrix map file for this
    location. 

YZ - the alleles given in the scan results, major first (if present,
     otherwise minor), then minor (if present, otherwise major). 

D - dbSNP ID for this location. 


* The "name" field in the exome sequencing results is given as X>Y:P:D
  (Example: "A>R:1:rs2761550").

X - the reference (hg18) data at this location 

Y - individual data at this location, in FASTA format ("R" = "G or A") 

P - PGP-ID (1..10) 

D - dbSNP ID for this location, if applicable. 


==Redacted data==

Exome sequencing results for PGP-6 were made using a subset of the available
reads. All reads for which alignments were found on chromosome 19 between
positions 48000000 and 52000000 were removed. Reads for which no alignments
were found were also removed. All Affymetrix SNPs for PGP6 on chromosome 19
between positions 48100000 and 52100000 were also removed.

==Exome sequencing process==

Samples were sequenced on an Illumina GA2--with one or more lanes per
sample--at http://genome.med.harvard.edu A total of two gigabases of "purity
filtered" sequence were generated for this release.  Exome data processing
was performed by maq 0.6.7 (map | mapmerge | assemble | cns2fq).  Single
end, 36bp reads were placed against ~55,000 capture regions w/coordinates on
HG18 in the "pgp#.cns.fq.txt.gz" files.  For all non-reference bp calls, a
minimum quality and coverage threshold was applied to raise the concordance
with Affymetrix data to 99%. More information will be made available with a
forthcoming publication.