Presented as poster #1397, ASHG Meeting, October 2000.

A simple, efficient procedure for mapping SNPs to genomic sequence.
P.S. Chines, K. Silander, N. Narisu, M.R. Erdos, A.K. Voltz2, K. Mohlke, 
F.S. Collins. DIR/GMBB, NHGRI, Bethesda, MD. 2DIR/IDRB, NHGRI, Baltimore,
MD.

Finding allelic association using a dense map of SNPs appears to be one of
the most promising approaches in mapping complex disease genes.  As of
June 5, 2000, over 135,000 SNPs and 87.8% of the human genomic sequence is
available in draft or finished form in NCBI databases.  But applying these
resources to generate dense SNP maps is not yet straightforward.  We
present a simple, fast method for finding SNPs that map to a region of
interest.

Method: We downloaded draft and finished sequence in a region of interest
in FASTA format.  We also downloaded SNP sequences and formatted them for
e-PCR, including information on SNP alleles and position within PCR
product.  Using a version of e-PCR modified to report which strand was
hit, we mapped the SNPs to the FASTA sequence. A Perl script calculated
the exact position of the SNP within each clone and extracted flanking
sequence.  From this list, we selected SNPs based on location, allele
frequency and confirmation status.  Finally, we BLASTed the flanking
sequences of selected SNPs against NCBI's HTGS database to identify
repetitive and non-specific sequences.

Result: We mapped 686 SNPs to a 16cM region of Chromosome 20q, using the
sequence of 83 clones available in draft or finished form (of 99 in the
minimal tiling path), while NCBI's website showed 1539 refSNPs mapped to
all of Chromosome 20, with 322 SNPs identifiably in our region.

Using e-PCR for initial screening, while reserving BLAST for evaluating
the uniqueness of a limited set of selected SNPs, makes this method both
faster and easier than using BLAST alone.  Because SNPs are mapped to
precise clone coordinates, duplicates are easily identified, and the SNP
sequence is confirmed by comparison with high-quality genomic sequence.
Arbitrarily long stretches of genomic sequence can be extracted for
designing SNP genotype assays.

The approach described here allows researchers to produce a dense SNP map
for a given region of interest based on the most up-to-date resources
available to supplement the pre-computed data from NCBI.  For information
about this software, contact the author at pchines@nhgri.nih.gov.