Tumor Microbiome Diversity and Composition Influence Pancreatic Cancer Outcomes.

Most patients diagnosed with resected pancreatic adenocarcinoma (PDAC) survive less than 5 years, but a minor subset survives longer. Here, we dissect the role of the tumor microbiota and the immune system in influencing long-term survival. Using 16S rRNA gene sequencing, we analyzed the tumor microbiome composition in PDAC patients with short-term survival (STS) and long-term survival (LTS). We found higher alpha-diversity in the tumor microbiome of LTS patients and identified an intra-tumoral microbiome signature (Pseudoxanthomonas-Streptomyces-Saccharopolyspora-Bacillus clausii) highly predictive of long-term survivorship in both discovery and validation cohorts. Through human-into-mice fecal microbiota transplantation (FMT) experiments from STS, LTS, or control donors, we were able to differentially modulate the tumor microbiome and affect tumor growth as well as tumor immune infiltration. Our study demonstrates that PDAC microbiome composition, which cross-talks to the gut microbiome, influences the host immune response and natural history of the disease.

An atlas of epithelial cell states and plasticity in lung adenocarcinoma.

Understanding the cellular processes that underlie early lung adenocarcinoma (LUAD) development is needed to devise intervention strategies1. Here we studied 246,102 single epithelial cells from 16 early-stage LUADs and 47 matched normal lung samples. Epithelial cells comprised diverse normal and cancer cell states, and diversity among cancer cells was strongly linked to LUAD-specific oncogenic drivers. KRAS mutant cancer cells showed distinct transcriptional features, reduced differentiation and low levels of aneuploidy. Non-malignant areas surrounding human LUAD samples were enriched with alveolar intermediate cells that displayed elevated KRT8 expression (termed KRT8+ alveolar intermediate cells (KACs) here), reduced differentiation, increased plasticity and driver KRAS mutations. Expression profiles of KACs were enriched in lung precancer cells and in LUAD cells and signified poor survival. In mice exposed to tobacco carcinogen, KACs emerged before lung tumours and persisted for months after cessation of carcinogen exposure. Moreover, they acquired Kras mutations and conveyed sensitivity to targeted KRAS inhibition in KAC-enriched organoids derived from alveolar type 2 (AT2) cells. Last, lineage-labelling of AT2 cells or KRT8+ cells following carcinogen exposure showed that KACs are possible intermediates in AT2-to-tumour cell transformation. This study provides new insights into epithelial cell states at the root of LUAD development, and such states could harbour potential targets for prevention or intervention.

Molecular mechanisms of the preventable causes of cancer in the United States.

Annually, there are 1.6 million new cases of cancer and nearly 600,000 cancer deaths in the United States alone. The public health burden associated with these numbers has motivated enormous research efforts into understanding the root causes of cancer. These efforts have led to the recognition that between 40% and 45% of cancers are associated with preventable risk factors and, importantly, have identified specific molecular mechanisms by which these exposures modify human physiology to induce or promote cancer. The increasingly refined knowledge of these mechanisms, which we summarize here, emphasizes the need for greater efforts toward primary cancer prevention through mitigation of modifiable risk factors. It also suggests exploitable avenues for improved secondary prevention (which includes the development of therapeutics designed for cancer interception and enhanced techniques for noninvasive screening and early detection) based on detailed knowledge of early neoplastic pathobiology. Such efforts would complement the current emphasis on the development of therapeutic approaches to treat established cancers and are likely to result in far greater gains in reducing morbidity and mortality.

Genomic Landscape of Lynch Syndrome Colorectal Neoplasia Identifies Shared Mutated Neoantigens for Immunoprevention.

BACKGROUND & AIMS: Lynch syndrome (LS) carriers develop mismatch repair-deficient neoplasia with high neoantigen (neoAg) rates. No detailed information on targetable neoAgs from LS precancers exists, which is crucial for vaccine development and immune-interception strategies. We report a focused somatic mutation and frameshift-neoAg landscape of microsatellite loci from colorectal polyps without malignant potential (PWOMP), precancers, and early-stage cancers in LS carriers. METHODS: We generated paired whole-exome and transcriptomic sequencing data from 8 colorectal PWOMP, 41 precancers, 8 advanced precancers, and 12 early-stage cancers of 43 LS carriers. A computational pipeline was developed to predict, rank, and prioritize the top 100 detected mutated neoAgs that were validated in vitro using ELISpot and tetramer assays. RESULTS: Mutation calling revealed >10 mut/Mb in 83% of cancers, 63% of advanced precancers, and 20% of precancers. Cancers displayed an average of 616 MHC-I neoAgs/sample, 294 in advanced precancers, and 107 in precancers. No neoAgs were detected in PWOMP. A total of 65% of our top 100 predicted neoAgs were immunogenic in vitro, and were present in 92% of cancers, 50% of advanced precancers, and 29% of precancers. We observed increased levels of naïve CD8+ and memory CD4+ T cells in mismatch repair-deficient cancers and precancers via transcriptomics analysis. CONCLUSIONS: Shared frameshift-neoAgs are generated within unstable microsatellite loci at initial stages of LS carcinogenesis and can induce T-cell responses, generating opportunities for vaccine development, targeting LS precancers and early-stage cancers.

Chromosomal imbalances detected via RNA-sequencing in 28 cancers.

MOTIVATION: RNA-sequencing (RNA-seq) of tumor tissue is typically only used to measure gene expression. Here, we present a statistical approach that leverages existing RNA-seq data to also detect somatic copy number alterations (SCNAs), a pervasive phenomenon in human cancers, without a need to sequence the corresponding DNA. RESULTS: We present an analysis of 4942 participant samples from 28 cancers in The Cancer Genome Atlas (TCGA), demonstrating robust detection of SCNAs from RNA-seq. Using genotype imputation and haplotype information, our RNA-based method had a median sensitivity of 85% to detect SCNAs defined by DNA analysis, at high specificity (∼95%). As an example of translational potential, we successfully replicated SCNA features associated with breast cancer subtypes. Our results credential haplotype-based inference based on RNA-seq to detect SCNAs in clinical and population-based settings. AVAILABILITY AND IMPLEMENTATION: The analyses presented use the data publicly available from TCGA Research Network (http://cancergenome.nih.gov/). See Methods for details regarding data downloads. hapLOHseq software is freely available under The MIT license and can be downloaded from http://scheet.org/software.html. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Causes of Clonal Hematopoiesis: a Review.

PURPOSE OF REVIEW: Clonal hematopoiesis (CH) is an age-dependent process detectable using advanced sequencing technologies and is associated with multiple adverse health outcomes including cardiovascular disease and cancer. The purpose of this review is to summarize known causes of CH mutations and to identify key areas and considerations for future research on CH. RECENT FINDINGS: Studies have identified multiple potential causes of CH mutations including smoking, cancer therapies, cardiometabolic disease, inflammation, and germline risk factors. Additionally, large-scale studies have facilitated the identification of gene-specific effects of CH mutation risk factors that may have unique downstream health implications. For example, cancer therapies and sources of environmental radiation appear to cause CH through their impact on DNA damage repair genes. There is a growing body of evidence defining risk factors for CH mutations. Standardization in the identification of CH mutations may have important implications for future research. Additional studies in underrepresented populations and their diverse environmental exposures are needed to facilitate broad public health impact of the study of CH mutations.

Association of clonal hematopoiesis mutations with clinical outcomes: A systematic review and meta-analysis.

Clonal hematopoiesis (CH) mutations are common among individuals without known hematologic disease. CH mutations have been associated with numerous adverse clinical outcomes across many different studies. We systematically reviewed the available literature for clinical outcomes associated with CH mutations in patients without hematologic disease. We searched PubMed, EMBASE, and Scopus for eligible studies. Three investigators independently extracted the data, and each study was verified by a second author. Risk of bias was assessed using the Newcastle-Ottawa Scale. We identified 32 studies with 56 cohorts that examine the association between CH mutations and clinical outcomes. We conducted meta-analyses comparing outcomes among individuals with and without detectable CH mutations. We conducted meta-analyses for cardiovascular diseases (nine studies; HR = 1.61, 95% CI = 1.26-2.07, p = .0002), hematologic malignancies (seven studies; HR = 5.59, 95% CI = 3.31-9.45, p < .0001), therapy-related myeloid neoplasms (four studies; HR = 7.55, 95% CI = 4.3-13.57, p < .001), and death (nine studies; HR = 1.34, 95% CI = 1.2-1.5, p < .0001). The cardiovascular disease analysis was further stratified by variant allele fraction (VAF) and gene, which showed a statistically significant association only with a VAF of ≥ 10% (HR = 1.42, 95% CI = 1.24-1.62, p < .0001), as well as statistically significant associations for each gene examined with the largest magnitude of effect found for CH mutations in JAK2 (HR = 3.5, 95% CI = 1.84-6.68, p < .0001). Analysis of the association of CH mutations with hematologic malignancy demonstrated a numeric stepwise increase in risk with increasing VAF thresholds. This analysis strongly supports the association of CH mutations with a clinically meaningful increased risk of adverse clinical outcomes among individuals without hematologic disease, particularly with increasing VAF thresholds.

Integrated case-control and somatic-germline interaction analyses of soft-tissue sarcoma.

PURPOSE: The contribution of rare genetic variation in the development of soft-tissue sarcoma (STS) remains underexplored. To address this gap, we conducted a whole-exome case-control and somatic-germline interaction study to identify and characterise STS susceptible genes. METHODS: The study involved 219 STS cases from The Cancer Genome Atlas and 3507 controls. All cases and controls were matched genetically onEuropean ancestry based on the 1000 Genomes project. Cross-platform technological stratification was performed with XPAT and gene-based association tests with VAAST 2. RESULTS: NF1 exhibited the strongest genome-wide signal across the six subtypes, with p=1×10-5. We also observed nominally significant association signals for three additional genes of interest, TP53 (p=0.0025), RB1 (p=0.0281), and MSH2 (p=0.0085). BAG1, which has not previously been implicated in STS, exhibited the strongest genome-wide signal after NF1, with p=6×10-5. The association signals for NF1 and MSH2 were driven primarily by truncating variants, with ORs of 39 (95% CI: 7.1 to 220) for NF1 and 33 (95% CI: 2.4 to 460) for MSH2. In contrast, the association signals for RB1 and BAG1 were driven primarily by predicted damaging missense variants, with estimated ORs of 12 (95% CI: 2.4 to 59) for RB1 and 20 (95% CI: 1.4 to 300) for BAG1. CONCLUSIONS: Our results confirm that pathogenic variants in NF1, RB1 and TP53 confer large increases in the risk of developing multiple STS subtypes, provide support for the role of MSH2 in STS susceptibility and identify BAG1 as a novel candidate STS risk gene.

Augmented Lipocalin-2 Is Associated with Chronic Obstructive Pulmonary Disease and Counteracts Lung Adenocarcinoma Development.

Rationale: Early pathogenesis of lung adenocarcinoma (LUAD) remains largely unknown. We found that, relative to wild-type littermates, the innate immunomodulator Lcn2 (lipocalin-2) was increased in normal airways from mice with knockout of the airway lineage gene Gprc5a (Gprc5a-/-) and that are prone to developing inflammation and LUAD. Yet, the role of LCN2 in lung inflammation and LUAD is poorly understood.Objectives: Delineate the role of Lcn2 induction in LUAD pathogenesis.Methods: Normal airway brushings, uninvolved lung tissues, and tumors from Gprc5a-/- mice before and after tobacco carcinogen exposure were analyzed by RNA sequencing. LCN2 mRNA was analyzed in public and in-house data sets of LUAD, lung squamous cancer (LUSC), chronic obstructive pulmonary disease (COPD), and LUAD/LUSC with COPD. LCN2 protein was immunohistochemically analyzed in a tissue microarray of 510 tumors. Temporal lung tumor development, gene expression programs, and host immune responses were compared between Gprc5a-/- and Gprc5a-/-/Lcn2-/- littermates.Measurements and Main Results:Lcn2 was progressively elevated during LUAD development and positively correlated with proinflammatory cytokines and inflammation gene sets. LCN2 was distinctively elevated in human LUADs, but not in LUSCs, relative to normal lungs and was associated with COPD among smokers and patients with LUAD. Relative to Gprc5a-/- mice, Gprc5a-/-/Lcn2-/- littermates exhibited significantly increased lung tumor development concomitant with reduced T-cell abundance (CD4+) and richness, attenuated antitumor immune gene programs, and increased immune cell expression of protumor inflammatory cytokines.Conclusions: Augmented LCN2 expression is a molecular feature of COPD-associated LUAD and counteracts LUAD development in vivo by maintaining antitumor immunity.

Embracing study heterogeneity for finding genetic interactions in large-scale research consortia.

Genetic interactions have been recognized as a potentially important contributor to the heritability of complex diseases. Nevertheless, due to small effect sizes and stringent multiple-testing correction, identifying genetic interactions in complex diseases is particularly challenging. To address the above challenges, many genomic research initiatives collaborate to form large-scale consortia and develop open access to enable sharing of genome-wide association study (GWAS) data. Despite the perceived benefits of data sharing from large consortia, a number of practical issues have arisen, such as privacy concerns on individual genomic information and heterogeneous data sources from distributed GWAS databases. In the context of large consortia, we demonstrate that the heterogeneously appearing marginal effects over distributed GWAS databases can offer new insights into genetic interactions for which conventional methods have had limited success. In this paper, we develop a novel two-stage testing procedure, named phylogenY-based effect-size tests for interactions using first 2 moments (YETI2), to detect genetic interactions through both pooled marginal effects, in terms of averaging site-specific marginal effects, and heterogeneity in marginal effects across sites, using a meta-analytic framework. YETI2 can not only be applied to large consortia without shared personal information but also can be used to leverage underlying heterogeneity in marginal effects to prioritize potential genetic interactions. We investigate the performance of YETI2 through simulation studies and apply YETI2 to bladder cancer data from dbGaP.

System for Quality-Assured Data Analysis: Flexible, reproducible scientific workflows.

The reproducibility of scientific processes is one of the paramount problems of bioinformatics, an engineering problem that must be addressed to perform good research. The System for Quality-Assured Data Analysis (SyQADA), described here, seeks to address reproducibility by managing many of the details of procedural bookkeeping in bioinformatics in as simple and transparent a manner as possible. SyQADA has been used by persons with backgrounds ranging from expert programmer to Unix novice, to perform and repeat dozens of diverse bioinformatics workflows on tens of thousands of samples, consuming over 80 CPU-months of computing on over 300,000 individual tasks of scores of projects on laptops, computer servers, and computing clusters. SyQADA is especially well-suited for paired-sample analyses found in cancer tumor-normal studies. SyQADA executable source code, documentation, tutorial examples, and workflows used in our lab is available from http://scheet.org/software.html.

Directional allelic imbalance profiling and visualization from multi-sample data with RECUR.

MOTIVATION: Genetic analysis of cancer regularly includes two or more samples from the same patient. Somatic copy number alterations leading to allelic imbalance (AI) play a critical role in cancer initiation and progression. Directional analysis and visualization of the alleles in imbalance in multi-sample settings allow for inference of recurrent mutations, providing insights into mutation rates, clonality and the genomic architecture and etiology of cancer. RESULTS: The REpeat Chromosomal changes Uncovered by Reflection (RECUR) is an R application for the comparative analysis of AI profiles derived from SNP array and next-generation sequencing data. The algorithm accepts genotype calls and 'B allele' frequencies (BAFs) from at least two samples derived from the same individual. For a predefined set of genomic regions with AI, RECUR compares BAF values among samples. In the presence of AI, the expected value of a BAF can shift in two possible directions, reflecting an increased or decreased abundance of the maternal haplotype, relative to the paternal. The phenomenon of opposite haplotype shifts, or 'mirrored subclonal allelic imbalance', is a form of heterogeneity, and has been linked to clinico-pathological features of cancer. RECUR detects such genomic segments of opposite haplotypes in imbalance and plots BAF values for all samples, using a two-color scheme for intuitive visualization. AVAILABILITY AND IMPLEMENTATION: RECUR is available as an R application. Source code and documentation are available at scheet.org. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Clonal evolution in breast cancer revealed by single nucleus genome sequencing.

Sequencing studies of breast tumour cohorts have identified many prevalent mutations, but provide limited insight into the genomic diversity within tumours. Here we developed a whole-genome and exome single cell sequencing approach called nuc-seq that uses G2/M nuclei to achieve 91% mean coverage breadth. We applied this method to sequence single normal and tumour nuclei from an oestrogen-receptor-positive (ER(+)) breast cancer and a triple-negative ductal carcinoma. In parallel, we performed single nuclei copy number profiling. Our data show that aneuploid rearrangements occurred early in tumour evolution and remained highly stable as the tumour masses clonally expanded. In contrast, point mutations evolved gradually, generating extensive clonal diversity. Using targeted single-molecule sequencing, many of the diverse mutations were shown to occur at low frequencies (<10%) in the tumour mass. Using mathematical modelling we found that the triple-negative tumour cells had an increased mutation rate (13.3×), whereas the ER(+) tumour cells did not. These findings have important implications for the diagnosis, therapeutic treatment and evolution of chemoresistance in breast cancer.

Driver Mutations in Normal Airway Epithelium Elucidate Spatiotemporal Resolution of Lung Cancer.

Rationale: Uninvolved normal-appearing airway epithelium has been shown to exhibit specific mutations characteristic of nearby non-small cell lung cancers (NSCLCs). Yet, its somatic mutational landscape in patients with early-stage NSCLC is unknown.Objectives: To comprehensively survey the somatic mutational architecture of the normal airway epithelium in patients with early-stage NSCLC.Methods: Multiregion normal airways, comprising tumor-adjacent small airways, tumor-distant large airways, nasal epithelium and uninvolved normal lung (collectively airway field), matched NSCLCs, and blood cells (n = 498) from 48 patients were interrogated for somatic single-nucleotide variants by deep-targeted DNA sequencing and for chromosomal allelic imbalance events by genome-wide genotype array profiling. Spatiotemporal relationships between the airway field and NSCLCs were assessed by phylogenetic analysis.Measurements and Main Results: Genomic airway field carcinogenesis was observed in 25 cases (52%). The airway field epithelium exhibited a total of 269 somatic mutations in most patients (n = 36) including key drivers that were shared with the NSCLCs. Allele frequencies of these acquired variants were overall higher in NSCLCs. Integrative analysis of single-nucleotide variants and allelic imbalance events revealed driver genes with shared "two-hit" alterations in the airway field (e.g., TP53, KRAS, KEAP1, STK11, and CDKN2A) and those with single hits progressing to two in the NSCLCs (e.g., PIK3CA and NOTCH1).Conclusions: Tumor-adjacent and tumor-distant normal-appearing airway epithelia exhibit somatic driver alterations that undergo selection-driven clonal expansion in NSCLC. These events offer spatiotemporal insights into the development of NSCLC and, thus, potential targets for early treatment.

XPAT: a toolkit to conduct cross-platform association studies with heterogeneous sequencing datasets.

High-throughput sequencing data are increasingly being made available to the research community for secondary analyses, providing new opportunities for large-scale association studies. However, heterogeneity in target capture and sequencing technologies often introduce strong technological stratification biases that overwhelm subtle signals of association in studies of complex traits. Here, we introduce the Cross-Platform Association Toolkit, XPAT, which provides a suite of tools designed to support and conduct large-scale association studies with heterogeneous sequencing datasets. XPAT includes tools to support cross-platform aware variant calling, quality control filtering, gene-based association testing and rare variant effect size estimation. To evaluate the performance of XPAT, we conducted case-control association studies for three diseases, including 783 breast cancer cases, 272 ovarian cancer cases, 205 Crohn disease cases and 3507 shared controls (including 1722 females) using sequencing data from multiple sources. XPAT greatly reduced Type I error inflation in the case-control analyses, while replicating many previously identified disease-gene associations. We also show that association tests conducted with XPAT using cross-platform data have comparable performance to tests using matched platform data. XPAT enables new association studies that combine existing sequencing datasets to identify genetic loci associated with common diseases and other complex traits.

Extensive Hidden Genomic Mosaicism Revealed in Normal Tissue.

Genomic mosaicism arising from post-zygotic mutation has recently been demonstrated to occur in normal tissue of individuals ascertained with varied phenotypes, indicating that detectable mosaicism may be less an exception than a rule in the general population. A challenge to comprehensive cataloging of mosaic mutations and their consequences is the presence of heterogeneous mixtures of cells, rendering low-frequency clones difficult to discern. Here we applied a computational method using estimated haplotypes to characterize mosaic megabase-scale structural mutations in 31,100 GWA study subjects. We provide in silico validation of 293 previously identified somatic mutations and identify an additional 794 novel mutations, most of which exist at lower aberrant cell fractions than have been demonstrated in previous surveys. These mutations occurred across the genome but in a nonrandom manner, and several chromosomes and loci showed unusual levels of mutation. Our analysis supports recent findings about the relationship between clonal mosaicism and old age. Finally, our results, in which we demonstrate a nearly 3-fold higher rate of clonal mosaicism, suggest that SNP-based population surveys of mosaic structural mutations should be conducted with haplotypes for optimal discovery.

The BE GONE trial study protocol: a randomized crossover dietary intervention of dry beans targeting the gut microbiome of overweight and obese patients with a history of colorectal polyps or cancer.

BACKGROUND: Mouse and human studies support the promise of dry beans to improve metabolic health and to lower cancer risk. In overweight/obese patients with a history of colorectal polyps or cancer, the Beans to Enrich the Gut microbiome vs. Obesity's Negative Effects (BE GONE) trial will test whether and how an increase in the consumption of pre-cooked, canned dry beans within the context of usual diet and lifestyle can enhance the gut landscape to improve metabolic health and reduce cancer risk. METHODS/DESIGN: This randomized crossover trial is designed to characterize changes in (1) host markers spanning lipid metabolism, inflammation, and obesity-related cancer risk; (2) compositional and functional profiles of the fecal microbiome; and (3) host and microbial metabolites. With each subject serving as their own control, the trial will compare the participant's usual diet with (intervention) and without (control) dry beans. Canned, pre-cooked dry beans are provided to participants and the usual diet continually assessed and monitored. Following a 4-week run-in and equilibration period, each participant provides a total of 5 fasting blood and 6 stool samples over a total period of 16 weeks. The intervention consists of a 2-week ramp-up of dry bean intake to 1 cup/d, which is then continued for an additional 6 weeks. Intra- and inter-individual outcomes are assessed across each crossover period with consideration of the joint or modifying effects of the usual diet and baseline microbiome. DISCUSSION: The BE GONE trial is evaluating a scalable dietary prevention strategy targeting the gut microbiome of high-risk patients to mitigate the metabolic and inflammatory effects of adiposity that influence colorectal cancer risk, recurrence, and survival. The overarching scientific goal is to further elucidate interactions between diet, the gut microbiome, and host metabolism. Improved understanding of the diet-microbiota interplay and effective means to target these relationships will be key to the future of clinical and public health approaches to cancer and other major diet- and obesity-related diseases. TRIAL REGISTRATION: This protocol is registered with the U.S. National Institutes of Health trial registry, ClinicalTrials.gov, under the identifier NCT02843425. First posted July 25, 2016; last verified January 25, 2019.

Strategies for identification of somatic variants using the Ion Torrent deep targeted sequencing platform.

BACKGROUND: 'Next-generation' (NGS) sequencing has wide application in medical genetics, including the detection of somatic variation in cancer. The Ion Torrent-based (IONT) platform is among NGS technologies employed in clinical, research and diagnostic settings. However, identifying mutations from IONT deep sequencing with high confidence has remained a challenge. We compared various computational variant-calling methods to derive a variant identification pipeline that may improve the molecular diagnostic and research utility of IONT. RESULTS: Using IONT, we surveyed variants from the 409-gene Comprehensive Cancer Panel in whole-section tumors, intra-tumoral biopsies and matched normal samples obtained from frozen tissues and blood from four early-stage non-small cell lung cancer (NSCLC) patients. We used MuTect, Varscan2, IONT's proprietary Ion Reporter, and a simple subtraction we called "Poor Man's Caller." Together these produced calls at 637 loci across all samples. Visual validation of 434 called variants was performed, and performance of the methods assessed individually and in combination. Of the subset of inspected putative variant calls (n=223) in genomic regions that were not intronic or intergenic, 68 variants (30%) were deemed valid after visual inspection. Among the individual methods, the Ion Reporter method offered perhaps the most reasonable tradeoffs. Ion Reporter captured 83% of all discovered variants; 50% of its variants were visually validated. Aggregating results from multiple packages offered varied improvements in performance. CONCLUSIONS: Overall, Ion Reporter offered the most attractive performance among the individual callers. This study suggests combined strategies to maximize sensitivity and positive predictive value in variant calling using IONT deep sequencing.

On meta- and mega-analyses for gene-environment interactions.

Gene-by-environment (G × E) interactions are important in explaining the missing heritability and understanding the causation of complex diseases, but a single, moderately sized study often has limited statistical power to detect such interactions. With the increasing need for integrating data and reporting results from multiple collaborative studies or sites, debate over choice between mega- versus meta-analysis continues. In principle, data from different sites can be integrated at the individual level into a "mega" data set, which can be fit by a joint "mega-analysis." Alternatively, analyses can be done at each site, and results across sites can be combined through a "meta-analysis" procedure without integrating individual level data across sites. Although mega-analysis has been advocated in several recent initiatives, meta-analysis has the advantages of simplicity and feasibility, and has recently led to several important findings in identifying main genetic effects. In this paper, we conducted empirical and simulation studies, using data from a G × E study of lung cancer, to compare the mega- and meta-analyses in four commonly used G × E analyses under the scenario that the number of studies is small and sample sizes of individual studies are relatively large. We compared the two data integration approaches in the context of fixed effect models and random effects models separately. Our investigations provide valuable insights in understanding the differences between mega- and meta-analyses in practice of combining small number of studies in identifying G × E interactions.