Single nucleotide polymorphisms (SNPs) are likely to contribute to the stud
y of complex genetic diseases. The genomic sequence of human chromosome 21q
was recently completed with 225 annotated genes, thus permitting efficient
identification and precise mapping of potential cSNPs by bioinformatics ap
proaches. Here we present a human chromosome 21 (HC21) cSNP database and th
e first chromosome-specific cSNP map. Potential cSNPs were generated using
three approaches: (1) Alignment of the complete HC21 genomic sequence to co
gnate ESTs and mRNAs. Candidate cSNPs were automatically extracted using a
novel program for context-dependent SNP identification that efficiently dis
criminates between true variation, poor quality sequencing, and paralogous
gene alignments. (2) Multiple alignment of all known HC21 genes to all othe
r human database entries. (3) Gene-targeted cSNP discovery. To dare we have
identified 377 cSNPs averaging -1 SNP per 1.5 kb of transcribed sequence,
covering 65% of known genes in the chromosome. Validation of our bioinforma
tics approach was demonstrated by a confirmation rate of 78% for the predic
ted cSNPs, and in total 32% of the cSNPs in our database have been confirme
d. The database is publicly available at http://csnp.unige.ch or http://csn
p.isb-sib.ch. These SNPs provide a tool to study the contribution of HC21 l
oci to complex diseases such as bipolar affective disorder and allele-speci
fic contributions to Down syndrome phenotypes.