Japanese genome-wide association study identifies a significant colorectal cancer susceptibility locus at chromosome 10p14

Abstract

Genome-wide association studies are a powerful tool for searching for disease susceptibility loci. Several studies identifying single nucleotide polymorphisms (SNP) connected intimately to the onset of colorectal cancer (CRC) have been published, but there are few reports of genome-wide association studies in Japan. To identify genetic variants that modify the risk of CRC oncogenesis, especially in the Japanese population, we performed a multi-stage genome-wide association study using a large number of samples: 1846 CRC cases and 2675 controls. We identified 4 SNP (rs7912831, rs4749812, rs7898455 and rs10905453) in chromosome region 10p14 associated with CRC; however, there are no coding or non-coding genes within this region of fairly extensive linkage disequilibrium (a 500-kb block) on 10p14. Our study revealed that the 10p14 locus is significantly correlated with susceptibility to CRC in the Japanese population, in accordance with the results of multiple studies in other races.

Keywords: 10p14; Colorectal cancer; Japanese; genome-wide association study; single nucleotide polymorphism.

Figure 1

The complete design of the genome‐wide association study. In phase 1, 577 patients with colorectal cancer (CRC) and 571 controls were genotyped for 500 568 single nucleotide polymorphisms (SNP) with Affymetrix 500 K chip sets. Two additional rounds of screening using the Illumina GoldenGate Assay (1536 SNP for phase 2.2) and TaqMan Assay (21 SNP for phase 3) were performed to identify significant SNP.

Figure 2

(a) Manhattan plots from the phase 1 genome‐wide association results. P‐values (−log10 P, y‐axis) are plotted against their respective chromosomal positions (x‐axis). Each chromosome is depicted in alternating blue and red. (b) Log quantile‐quantile P‐values between the expected (theoretical) P‐value and the observed P‐value. The genomic inflation factor (based on median χ²) is 1.10. If we set the odds ratio of the CRC‐related genotype as 1.4 and the allelic frequency in the control as 0.2, it will be located at the 36th quantile by the P‐value distribution. If we set the odds ratio of the CRC‐related genotype as 1.4 and the allelic frequency in the control as 0.4, it will be located at the 350th. If we set the odds ratio of the CRC‐related genotype as 1.2 and the allelic frequency in the control as 0.2, it will be located in greater than the 1000th, and the genotype will be difficult to determine.

Figure 3

Linkage disequilibrium (LD) structure at 10p14. The polymorphic site rs7912831 is depicted in an LD block of 500 kb where there are coding and non‐coding genes, such as non‐coding RNA and micro RNA. The SNP rs10795668, which has been previously reported by Tomlinson et al., is located close to rs7912831. Data were analyzed using Haploview software (v3.2)