Analysis of quantitative trait loci (eQTLs) in pancreatic cancer by RNA-sequencing

Tuesday, October 09, 2012 — Poster Session I
1:00 p.m. – 3:00 p.m	Natcher Conference Center, Building 45	NCI	BIOINFO-14

Authors

• H. Parikh
• J. Jia
• W. Xiao
• I. Collins
• J. Hoskins
• J. Powell
• S. Thorgeirsson
• J. Shi
• G. Petersen
• L. Amundadottir

Abstract

We are conducting expression quantitative trait loci (eQTLs) analysis in normal and tumor derived pancreatic tissue samples. We performed massively parallel sequencing to catalog the transcriptome of 50 normal or tumor derived human pancreatic tissues. Reads were mapped to the RefSeq database using the BWA software. Gene expression was calculated as tag counts. The EDASeq package was used to perform full-quantile normalization to adjust for differences between lanes. Blood derived DNA from the same subjects was genotyped using either the Illumina Omni1 or OmniExpress chips. We selected genes within 2 MB up- and downstream of the most significant SNPs in each of the four regions from the GWAS and analyzed SNPs in a region from 500 kb up- and downstream of the four genomic regions on chromosomes 1q32.1, 5p15.33, 9q34.2 and 13q22.1 associated with pancreatic cancer risk. We performed linear regression analysis between gene tag counts and SNP genotypes after correcting for age and sex. We identified 125 cis-eQTLs for chromosome 1q32.1, 2,757 cis-eQTLs for chromosome 5p15.33, 3,999 cis-eQTLs for chromosome 9q34.2 and 629 cis-eQTLs for chromosome13q22.1 with nominal P-value < 0.05. Together these results offer a unique picture of the functional aspects of pancreatic cancer risk loci.

back to top