Increasing the Number of SNP loci does not Necessarily Improve Prediction Power at Least in the Comparison of MTHFR SNP and Haplotypes

BACKGROUND: Rapid advances in genotyping technology have made it possible to easily utilize a large number of genetic markers. According to information theory, an increase in the number of markers provides more information; however, the clinical usefulness does not increase linearly. This study aimed to assess the effect of folic acid supplementation quantitatively in MTHFR haplotypes, and compare its prediction power with that of the C677T single nucleotide polymorphism (SNP) alone. METHODS: The study was a randomized, double-blind, placebo-controlled trial, designed in accordance with the CONSORT statement. The participants were 202 healthy Japanese males who were administered either folic acid at 1 mg/day or a placebo postoperatively for 3 months. The primary endpoint was the total plasma homocysteine levels (tHcy). Stratified analysis by HapMap-based tag SNPs was performed. RESULTS: Of 52 SNPs on the MTHFR gene, 4 SNP loci covering more than 80% of the information were selected, and the haplotypes were estimated. The haplotypes were classified into 3 groups (Hap0, Hap1, and Hap2), on the basis of the number of times the most frequent haplotype was present. The greatest decrease was observed in Hap2 (6.61 µmol/L), compared with the other haplotypes (Hap0, 2.67; Hap1, 2.60) (trend test, P < 0.01). The haplotype information obtained was not more informative than that obtained with grouping by a single SNP, C677T, which strongly influences enzyme activity. CONCLUSIONS: Grouping by the C677T SNP alone was almost as good a predictor of the homocysteine-lowering effects as was grouping by the 4 best SNPs. This shows that increasing the number of typed SNPs does not necessarily provide more information, at least for this gene. A more efficient, cost-informative method for analyzing genomic data is required.