| SNP Discovery - prediction and
verification Statistical analysis of all
major SNP databases has enabled us to develop, distribute, and represent all
of those databases. This enables us to target particular SNPs in
candidate genes for direct analysis within the Reynold's population.
Directed search
for predicted point mutations within gene coding regions
A
pilot study to experimentally test the effectiveness of nonsynonymous cSNP
(coding region SNP) prediction has been initiated as part of UTSW PGA
project.. Previously we showed that given both the frequency and location
of SNPs triggering human genetic disease are determined in part by local DNA
sequence environment (1, 2).
Public cSNP collections can therefore be mined to pinpoint DNA sequence
signatures statistically prone to point mutation. These methods
predict sites prone to mutation but cannot make any statements about the
expected allele frequency of cSNPs given that both natural selection and
population history determine allele penetrance.
However, if implicated bases are
indeed mutation-prone, then these alleles are expected to be found in a
sufficiently large, but accessible, random population. To this end,
nonsynonymous cSNPs in 20 genes implicated in cardiac disease were predicted
using database-derived BLOSUM-style scoring matrices and queried in the
Reynold’s Center cohort of 3,554 individuals according to the following
protocol:
Such a directed
search for as-of-yet undiscovered point mutations based on analysis of
mutation database trends may present one solution to finding rare alleles
contributing to multigenic diseases.
Current genotyping data and algorithm evaluation progress:
Batch 1 Progress: Fall 2003
Batch 2 Progress: Spring 2004
(Sequenome genotyping in progress)
References:
(1)
Gene.
2003 Jul 17;312:197-206.
(2)
Supplementary data on prediction methodology
|
