Pooled heteronuclear RNA (hnRNA) sequencing: a new tool for large-scale cis-acting regulatory haplotype (rHaplotype) discovery
1Scott Gurd, 1Kevin Lam, 1Bing Ge, 1Nicolas Martin, 1Eef Harmsen, 1Tony Kwan, 1,2Jacek Majewski, 1,2Tomi Pastinen
1McGill University and Genome Quebec Innovation Centre, Montreal, Quebec, Canada, 2Department of Human Genetics, McGill University, Montreal, Quebec, Canada
Common haplotypes associated with complex human disease traits are uncovered at increasing rates and in many cases no coding SNPs can be identified. Concurrent studies have also shown that population variation in human gene expression is partly governed by sequence variants in cis-acting rHaplotypes serving as functional candidates for disease association studies.
Approaches for identification of rHaplotypes include mapping of gene expression levels to genotypes. We have discovered cis-acting polymorphisms by direct measurements of allelic expression of the two alleles of each gene in human cell lines or tissues. These measurements are more sensitive and specific for cis-acting variation as compared to methods based on comparing expression differences between individuals, but due to low throughput are best suited to validation and fine-mapping of rHaplotypes.
We have now implemented a sequencing-based quantitation of relative differences of allele frequencies between DNA and RNA pools from same population of individuals, which identifies polymorphisms that are in linkage disequilibrium with regulatory variants. This screening coupled with validation and rHaplotype mapping by allelic expression provides a rapid and sensitive tool for discovery of regulatory variants. We have focused our screening to (un-spliced) hnRNA allowing specific interrogation of common haplotypes in the vicinity of human genes and focus on transcriptional cis-regulatory effects. Our early results in over 500 SNPs indicate that the method has exquisite sensitivity for cis-acting effects based on comparison to results from three large-scale expression profiling studies using different microarray platforms (Illumina and Affymetrix Focus or Exon arrays)in the same population of lymphoblastoid cell lines (HapMap CEU lines). In addition to picking up virtually all cis-effects identified by the expression profiling methods the pooled hnRNA sequencing approach led to identification many novel cis-acting effects, such as a strong association of INSIG2 allelic expression to a rHaplotype. This association was weak or undetectable by the microarray-based population expression studies. We further validated the association by allelic expression studies in an independent Caucasian lymphoblastoid cell panel. Our method can be easily applied to new ultra-hightroughput sequencing platforms allowing sensitive surveys of subtle cis-acting effects in population-based samples.
This work is supported by Genome Quebec and Genome Canada.