Variations in the DNA sequences can affect how we
develop diseases and respond to pathogens, chemicals, drugs, vaccines, and
other agents. The DNA sequence is read and compared between a normal and a
diseased person and the change in sequence of the amino acid A, T, C, or G is
noted. This change from the normal is called as single nucleotide
polymorphisms (SNPs). SNPs are usually considered to be point mutations.
A recent study conducted by Dr. Shenying Fang and colleagues of the University of Texas M. D. Anderson Cancer Center looked into the use of single nucleotide polymorphisms (SNPs)
to predict psoriasis. To achieve this, two issues were to be addressed. The first one being, what classification method should be used? For this, the authors proposed to use a classification method called sequential information bottleneck (sIB) to predict the occurrence of disease. The second issue was how to select a subset of SNPs from half millions of available SNPs.
For the study, a total of 2,798 samples and 451,724
SNPs were taken. The actual process for searching a set of SNPs which can
predict the susceptibility for psoriasis consisted of two steps. The first one
was to search top most 1,000 SNPs with high accuracy for prediction of
psoriasis from GWAS dataset. The second one was to look for an optimal SNP
subset for psoriasis prediction. The classical linear discriminant
analysis(LDA) method was compared with sequential information bottleneck (sIB)
for classification performance.
The best test harmonic mean of sensitivity and
specificity for predicting psoriasis by sIB was 0.674(95% CI: 0.650-0.698),
while only 0.520(95% CI: 0.472-0.524) was reported for predicting disease by
Study author Dr. Fnag concluded by saying that, "Our
classification methods achieved high prediction accuracy in this study,
determining the statistical significance of those models requires more
cost-effective methods or efficient computing system, neither of which is
available currently in our genome-wide study."
Result of the study indicates that a small set of SNPs
can predict disease status with average accuracy of 68% and makes it possible
to use SNP data for psoriasis prediction.