Mathieu Lemire

Genome-wide association studies have identified 55 genetic variants associated with colorectal cancer risk to date. However, potential causal genes and pathways regulated by these risk variants remain to be characterized. Therefore, we performed gene ontology enrichment and pathway analyses to determine if there was an enrichment of genes in proximity to the colorectal cancer risk variants that could further elucidate the probable causal genes and pathways involved in colorectal cancer biology. For the 65 unique genes that either contained, or were immediately neighboring up- and downstream, of these variants there was a significant enrichment for the KEGG pathway, Pathways in Cancer (p-value = 2.67 × 10−5) and an enrichment for multiple biological processes (FDR < 0.05), such as cell junction organization, tissue morphogenesis, regulation of SMAD protein phosphorylation, and odontogenesis identified through Gene Ontology analysis. To identify potential causal genes, we conducted a cis-expression quantitative trait loci (cis-eQTL) analysis using gene expression and genotype data from the Genotype-Tissue Expression (GTEx) Project portal in normal sigmoid (n = 124) and transverse (n = 169) colon tissue. In addition, we also did a cis-eQTL analysis on colorectal tumor tissue (n = 147) from The Cancer Genome Atlas (TCGA). We identified two risk alleles that were significant cis-eQTLs for FADS2 (rs1535) and COLCA1 and 2 (rs3802842) genes in the normal transverse colon tissue and two risk alleles that were significant cis-eQTLs for the CABLES2 (rs2427308) and LIPG (rs7229639) genes in the normal sigmoid colon tissue, but not tumor tissue. Our data reaffirm the potential to identify an enrichment for biological processes and candidate causal genes based on expression profiles correlated with genetic risk alleles of colorectal cancer, however, the identification of these significant cis-eQTLs is context and tissue specific.

Colorectal cancers (CRCs) undergo distinct genetic and epigenetic alterations. Expression of mutL homolog 1 (MLH1), a mismatch repair gene that corrects DNA replication errors, is lost in up to 15% of sporadic tumours due to mutation or, more commonly, due to DNA methylation of its promoter CpG island. A single nucleotide polymorphism (SNP) in the CpG island of MLH1 (MLH1-93G>A or rs1800734) is associated with CpG island hypermethylation and decreased MLH1 expression in CRC tumours. Further, in peripheral blood mononuclear cell (PBMC) DNA of both CRC cases and non-cancer controls, the variant allele of rs1800734 is associated with hypomethylation at the MLH1 shore, a region upstream of its CpG island that is less dense in CpG sites. To determine whether this genotype-epigenotype association is present in other tissue types, including colorectal tumours, we assessed DNA methylation in matched normal colorectal tissue, tumour, and PBMC DNA from 349 population-based CRC cases recruited from the Ontario Familial Colorectal Cancer Registry. Using the semi-quantitative real-time PCR-based MethyLight assay, MLH1 shore methylation was significantly higher in tumour tissue than normal colon or PBMCs (P < 0.01). When shore methylation levels were stratified by SNP genotype, normal colorectal DNA and PBMC DNA were significantly hypomethylated in association with variant SNP genotype (P < 0.05). However, this association was lost in tumour DNA. Among distinct stages of CRC, metastatic stage IV CRC tumours incurred significant hypomethylation compared to stage I–III cases, irrespective of genotype status. Shore methylation of MLH1 was not associated with MSI status or promoter CpG island hypermethylation, regardless of genotype. To confirm these results, bisulfite sequencing was performed in matched tumour and normal colorectal specimens from six CRC cases, including two cases per genotype (wildtype, heterozygous, and homozygous variant). Bisulfite sequencing results corroborated the methylation patterns found by MethyLight, with significant hypomethylation in normal colorectal tissue of variant SNP allele carriers. These results indicate that the normal tissue types tested (colorectum and PBMC) experience dynamic genotype-associated epigenetic alterations at the MLH1 shore, whereas tumour DNA incurs aberrant hypermethylation compared to normal DNA.

Over 50 loci associated with colorectal cancer (CRC) have been uncovered by genome-wide association studies (GWAS). Identifying additional loci has the potential to help elucidate aspects of the underlying biological processes leading to better understanding of the pathogenesis of the disease. We re-evaluated a GWAS by excluding controls that have family history of CRC or personal history of colorectal polyps, as we hypothesized that their inclusion reduces power to detect associations. This is supported empirically and through simulations. Two-phase GWAS analysis was performed in a total of 16,517 cases and 14,487 controls. We identified rs17094983, a SNP associated with risk of CRC [p = 2.5 × 10⁻¹⁰; odds ratio estimated by re-including all controls (OR) = 0.87, 95 % confidence interval (CI) 0.83–0.91; minor allele frequency (MAF) = 13 %]. Results were replicated in samples of African descent (1894 cases and 4703 controls; p = 0.01; OR = 0.86, 95 % CI 0.77–0.97; MAF = 16 %). Gene expression data in 195 colon adenocarcinomas and 59 normal colon tissues from two different studies revealed that this locus has genotypes that are associated with RTN1 (Reticulon 1) expression (p = 0.001), a protein-coding gene involved in survival and proliferation of cancer cells which is highly expressed in normal colon tissues but has significantly reduced expression in tumor cells (p = 1.3 × 10⁻⁸).

The interplay between genetic and epigenetic variation is only partially understood. One form of epigenetic variation is methylation at CpG sites, which can be measured as methylation quantitative trait loci (meQTL). Here we report that in a panel of lymphocytes from 1,748 individuals, methylation levels at 1,919 CpG sites are correlated with at least one distal (trans) single-nucleotide polymorphism (SNP) (P−13; FDR1 Mb apart. Over 90% of these pairs are replicated (FDRtrans-meQTLs are significantly enriched (P<0.001) for long non-coding transcripts (2.2-fold), known epigenetic regulators (2.3-fold), piwi-interacting RNA clusters (3.6-fold) and curated transcription factors (4.1-fold), including zinc-finger proteins (8.75-fold). Long-range epigenetic networks uncovered by this approach may be relevant to normal and disease states.

We propose an extension to quantile normalization that removes unwanted technical variation using control probes. We adapt our algorithm, functional normalization, to the Illumina 450k methylation array and address the open problem of normalizing methylation data with global epigenetic changes, such as human cancers. Using data sets from The Cancer Genome Atlas and a large case–control study, we show that our algorithm outperforms all existing normalization methods with respect to replication of results between experiments, and yields robust results even in the presence of batch effects. Functional normalization can be applied to any microarray platform, provided suitable control probes are available.

We aim to identify rare variants that have large effects on trait variance using a cost-efficient strategy. We use an oligogenic segregation analysis as a prioritizing tool for whole-exome sequencing studies to identify families more likely to harbor rare variants, by estimating the mean number of quantitative trait loci (QTLs) in each family. We hypothesize that families with additional QTLs, relative to the other families, are more likely to segregate functional rare variants. We test the association of rare variants with the traits only in regions where at least modest evidence of linkage with the trait is observed, thereby reducing the number of tests performed. We found that family 7 harbored an estimated two, one, and zero additional QTLs for traits Q1, Q2, and Q4, respectively. Two rare variants (C4S4935 and C6S2981) segregating in family 7 were associated with Q1 and explained a substantial proportion of the observed linkage signal. These rare variants have 31 and 22 carriers, respectively, in the 128-member family and entered through a single but different founder. For Q2, we found one rare variant unique to family 7 that showed small effect and weak evidence of association; this was a false positive. These results are a proof of principle that prioritizing the sequencing of carefully selected extended families is a simple and cost-efficient design strategy for sequencing studies aiming at identifying functional rare variants.

Using single-nucleotide polymorphisms (SNPs), we sought to predict classical class I and class II human leukocyte antigen (HLA) alleles, and test for their associations with rheumatoid arthritis (RA) in the North American Rheumatoid Arthritis Consortium sample of cases and controls, genotyped on the Illumina HumanHap550 BeadChip. We use publicly available databases of SNP data and HLA data to find SNPs or SNP-haplotypes to be used as surrogates for each HLA allele. To reduce the confounding effects of linkage disequilibrium with the HLA-DRB1 locus, we tested for the association conditional on the presence or absence of a shared epitope allele on the same haplotype as the target HLA allele. Using SNP surrogates, we find that components of the DQ8 serotype (DQA1*0301:DQB1*0302) are associated with RA, irrespective of the presence or absence of a shared epitope allele on their respective haplotypes. Knowledge of the haplotype structure in the HLA region is still necessary for better interpretation of the results.

In high-dimensional studies such as genome-wide association studies, the correction for multiple testing in order to control total type I error results in decreased power to detect modest effects. We present a new analytical approach based on the higher criticism statistic that allows identification of the presence of modest effects. We apply our method to the genome-wide study of rheumatoid arthritis provided in the Genetic Analysis Workshop 16 Problem 1 data set. There is evidence for unknown bias in this study that could be explained by the presence of undetected modest effects. We compared the asymptotic and empirical thresholds for the higher criticism statistic. Using the asymptotic threshold we detected the presence of modest effects genome-wide. We also detected modest effects using 90th percentile of the empirical null distribution as a threshold; however, there is no such evidence when the 95th and 99th percentiles were used. While the higher criticism method suggests that there is some evidence for modest effects, interpreting individual single-nucleotide polymorphisms with significant higher criticism statistics is of undermined value. The goal of higher criticism is to alert the researcher that genetic effects remain to be discovered and to promote the use of more targeted and powerful studies to detect the remaining effects.

Background: Genetic variants at the vitamin D receptor (VDR) locus are associated with asthma and atopy. We hypothesized that polymorphisms in other genes of the vitamin D pathway are associated with asthma or atopy. Methods: Eleven candidate genes were chosen for this study, five of which code for proteins in the vitamin D metabolism pathway (CYP27A1, CYP27B1, CYP2R1, CYP24A1, GC) and six that are known to be transcriptionally regulated by vitamin D (IL10, IL1RL1, CD28, CD86, IL8, SKIIP). For each gene, we selected a maximally informative set of common SNPs (tagSNPs) using the European-derived (CEU) HapMap dataset. A total of 87 SNPs were genotyped in a French-Canadian family sample ascertained through asthmatic probands (388 nuclear families, 1064 individuals) and evaluated using the Family Based Association Test (FBAT) program. We then sought to replicate the positive findings in four independent samples: two from Western Canada, one from Australia and one from the USA (CAMP). Results: A number of SNPs in the IL10, CYP24A1, CYP2R1, IL1RL1 and CD86 genes were modestly associated with asthma and atopy (p < 0.05). Two-gene models testing for both main effects and the interaction were then performed using conditional logistic regression. Two-gene models implicating functional variants in the IL10 and VDR genes as well as in the IL10 and IL1RL1 genes were associated with asthma (p < 0.0002). In the replicate samples, SNPs in the IL10 and CYP24A1 genes were again modestly associated with asthma and atopy (p < 0.05). However, the SNPs or the orientation of the risk alleles were different between populations. A two-gene model involving IL10 and VDR was replicated in CAMP, but not in the other populations. Conclusion: A number of genes involved in the vitamin D pathway demonstrate modest levels of association with asthma and atopy. Multilocus models testing genes in the same pathway are potentially more effective to evaluate the risk of asthma, but the effects are not uniform across populations.