07-0289 864..865 Editorial Genetic Association Studies of Cancer: Where Do We Go from Here? Timothy R. Rebbeck,1 Muin J. Khoury,2 and John D. Potter3 1Abramson Cancer Center, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; 2National Office of Public Health Genomics, Centers for Disease Control and Prevention, Atlanta, Georgia; and 3Fred Hutchinson Cancer Research Center, Seattle, Washington ‘‘A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines.’’ —Ralph Waldo Emerson A goal of genetic association studies is to identify disease- causing genes. True positive associations between genetic variants and disease traits have the potential both to elucidate biologic principles and to lead to translation of research into clinical or preventive practice. For instance, an environmental effect on cancer risk may vary with genotype; or response to a treatment can differ based on information contained in the genotype. However, the reputation of genetic association studies has suffered because of the difficulty identifying reproducible and biologically interpretable results. Although criteria for identifying truly causative events in disease etiology have been proposed in numerous settings (1-7), we need to advance both conception and methods to improve the execution, interpretation, and integration of genetic association studies. Studies on genetic associations in cancer have been published at an increasing pace. The CDC-curated published literature database on genetic associations (8) reveals the tremendous growth of this literature (Table 1). The annual number of reports of genetic associations involving cancer as an outcome has more than tripled over the past 6 years. These publications include an increasing number of articles on gene- gene and gene-environment interactions, as well as systematic reviews and meta-analyses that integrate the literature on genetic variants with that on cancer risk. This upward trend is likely to continue over the next few years, particularly with the advent of genome-wide association platforms in which hundreds of thousands of genetic variants will be studied simultaneously in large-scale consortium studies. So where is the field going? And how can CEBP help to bring order to this increasing flood of literature? A fundamental conceptual requirement for molecular epidemiology has been the insistence that findings of a genetic association study be replicable. Nonetheless, critics of molec- ular epidemiologic associations note, sometimes properly, that many such studies are not replicated (9); inconsistencies are regularly explained, at least in part, by inadequate approaches to study design and analysis. Recent data suggest that population genetics structure, such as deviation from Hardy- Weinberg equilibrium or linkage disequilibrium between the SNP under study and a causative SNP (10, 11), can affect association study results. Thus, some inconsistent associations can be explained by a more extensive consideration of the genomic or environmental context of the variants under study. Scientists are appropriately wary of incorrect inferences because they can lead subsequent research in the wrong direction, wasting valuable resources and time. Some journals have suggested that studies cannot be published unless evidence of replication is provided at the time of the initial publication (7).Taking this even further, othershave suggested that the lack of consistency in associations raises serious concerns about the value of molecular epidemiology as a discipline. Nonetheless, many studies are replicated (9), and many unreplicated associations could ultimately have biological or clinical relevance. Thus, a simple-minded requirement for replicating studies of genetic associations is in danger of becoming the hobgoblin of human genetics research. Overem- phasis on the apparent inconsistency of association-study results fails to acknowledge the complexity and heterogeneity that underlie all common diseases, including cancer. For example, an association may genuinely be present (or detectable) in some populations or subpopulations, but not others. An association may be observed only in a specific context (e.g., under exposure to a relevant agent), but such contexts are not always understood nor are the relevant expo- sures well measured. Complexity and heterogeneity are always ignored in the design, analysis, execution, or interpre- tation of association studies. Most recognize how unlikely it is that a single nucleotide variant (no matter how functionally relevant) has a simple relationship with the risk of developing complex diseases (12). It is far more likely that many genetic variants, in conjunction with environmental exposures, will explain variation in disease etiology, drug response, and prognosis. A correlate of this higher-order com- plexity is that the marginal effects of single genetic variants may be obscured, and studies that do not properly account for this complexity are likely to miss relevant etiologic events. Therefore, studies need to be designed, executed, and inter- preted in recognition of this complexity. A lack of consistency of associations may not always reflect an absence of causative association in disease but may instead be the consequence of true heterogeneity in the effects of genetic variants, exposures, or their interactions. To avoid having overly simplistic notions of consistency overtake and hamper scientific progress, a sophisticated framework is required to identify biologically or clinically meaningful genetic associations. The Human Genome Epide- miology (HuGENet) Network4 and Cancer Epidemiology, Bio- markers & Prevention (CEBP) have provided separate forums 864 4 http://www.cdc.gov/genomics/hugenet/default.htm Cancer Epidemiol Biomarkers Prev 2007;16(5). May 2007 Cancer Epidemiol Biomarkers Prev 2007;16(5):864–5 Received 3/28/07; accepted 3/28/07. Grant support: The Robert Wood Foundation, the Cancer Center Support Grant, National Cancer Institutegrant 5P30CA016672, andNationalCancer Institutegrant 5R01CA097893-02. Requests for reprints: Timothy Rebbeck, Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, 904 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021. Phone: 215-898-1793; Fax: 215-4409355. E-mail: trebbeck@cceb.med.upenn.edu Copyright D 2007 American Association for Cancer Research. doi:10.1158/1055-9965.EPI-07-0289 on April 5, 2021. © 2007 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from http://cebp.aacrjournals.org/ for the publication, interpretation, and synthesis of association studies. Now CEBP and HuGENet will team up to make sense of this evolving field. CEBP has been a leader in publishing genetic association studies in cancer but recognizes the limitations of the literature in attempting to identify truly causative associations with biologic, clinical, or public health relevance. HuGENet is an open global collaboration of individuals and groups interested in advancing the integration of knowledge and applications of genetic information in medicine and public health. HuGENet sponsors methodologic workshops, hosts a published literature database, promotes systematic reviews of genetic associations-HuGE reviews (13), and convenes collaborations among consortia and networks through its ‘‘network of networks’’ (14). CEBP and HuGENet have individually and jointly proposed guidelines bywhich an improved environment for the publication of association studies can be achieved (2, 15-17). These guidelines encompass biological plausibility; populationgenetics structure (including factors related to race or ethnicity); appropriate laboratory and questionnaire measurements of exposures; consideration of etiologic complexity and heterogeneity; appropriate study design; and the correct inferential methods that consider etiologic complexity. Current guidelines, however, remain inadequate and require further development. HuGENet serves as a focus for this effort and provides continuing guidance and direction for the field (18, 19). CEBP recognizes the merit of the structured approach that HuGENet can provide to evaluate association study results. CEBP therefore encourages the publication of methodological research that enhances the field of association studies, and applied association studies that appropriately consider etiologic complexity and replication, as well as meta- analyses that summarize the existing literature. CEBP further encourages authors interested in integrating information on genetic associations to use the HuGENet handbook for systematic reviews as a framework for providing meaningful information to the scientific literature (20). More than ever, as the pace of publication continues to increase, this field needs a collective effort to establishandapply thebest set of guidelines for the analysis, interpretation, and synthesis of genetic associations. References 1. WeissST.Association studies inasthmagenetics.AmJRespirCritCareMed 2001;164:2014–5. 2. Ioannidis JP. Commentary: grading the credibility of molecular evidence for complex diseases. Int J Epidemiol 2006;35:572–8; discussion 593–6. 3. Cooper DN, Nussbaum RL, Krawczak M. Proposed guidelines for papers describing DNA polymorphism-disease associations. Hum Genet 2002;110: 207–8. 4. Huizinga TW, Pisetsky DS, Kimberly RP. Associations, populations, and the truth: recommendations for genetic association studies in Arthritis & Rheumatism. Arthritis Rheum 2004;50:2066–71. 5. Rebbeck TR, Spitz M, Wu X. Assessing the function of genetic variants in candidate gene association studies. Nat Rev Genet 2004;5:589–97. 6. Freimer NB, Sabatti C. Guidelines for association studies in Human Molecular Genetics. Hum Mol Genet 2005;14:2481–3. 7. Framework for a fully powered risk engine. Nat Genet 2005;37:1153. 8. LinBK,ClyneM,WalshM,et al. Tracking theepidemiologyofhumangenes in the literature: the HuGE Published Literature database. Am J Epidemiol 2006;164:1–4. 9. Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschhorn JN. Meta- analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 2003;33:177–82. 10. Lin PI, Vance JM, Pericak-Vance MA, Martin ER. No gene is an island: the flip-flop phenomenon. Am J Hum Genet 2007;80:531–8. 11. Wittke-Thompson JK, Pluzhnikov A, Cox NJ. Rational inferences about departures from Hardy-Weinberg equilibrium. Am J Hum Genet 2005;76: 967–86. 12. Sing CF, Haviland MB, Reilly SL. Genetic architecture of common multi- factorial diseases. Ciba Found Symp 1996;197:211–29; discussion 229–32. 13. Little J, Khoury MJ, Bradley L, et al. The human genome project is complete. How do we develop a handle for the pump? Am J Epidemiol 2003;157:667–73. 14. Ioannidis JP, Gwinn M, Little J, et al. A road map for efficient and reliable human genome epidemiology. Nat Genet 2006;38:3–5. 15. Rebbeck TR, Martinez ME, Sellers TA, Shields PG, Wild CP, Potter JD. Genetic variation and cancer: improving the environment for publication of association studies [see comment]. Cancer Epidemiol Biomarkers Prev 2004; 13:1985–6. 16. Ioannidis JP. Journals should publishall ‘‘null’’ results and should sparingly publish ‘‘positive’’ results. Cancer Epidemiol Biomarkers Prev 2006;15:186. 17. Rebbeck TR, Ambrosone CB, Bell DA, et al. SNPs, haplotypes, and cancer: applications in molecular epidemiology. Cancer Epidemiol Biomarkers Prev 2004;13:681–7. 18. Seminara D, Khoury MJ, O’Brien TR, et al. The emergence of networks in human genome epidemiology: challenges and opportunities. Epidemiology 2007;18:1–8. 19. Ioannidis JP, Bernstein J, Boffetta P, et al. A network of investigator networks in human genome epidemiology. Am J Epidemiol 2005;162:302–4. 20. Little J, Higgins J. HuGENet HuGE review handbook, version 1.0. c2006 [cited 2007 Mar 9]. Available from: http://www.genesens.net/_intranet/ doc_nouvelles/HuGE%20Review%20Handbook%20v11.pdf. Table 1. Trends in the published literature on genetic associations in cancer, 2001-2006 Type of publication* Genetic associations Gene-gene and gene-environment interactions Meta-analyses and HuGE reviews Total 2001 333 139 5 383 2002 497 175 17 579 2003 555 165 14 644 2004 699 210 21 797 2005 983 319 33 1,090 2006 1,095 375 38 1,219 NOTE: Data are from analysis of the CDC HuGE Published Literature Database on February 12, 2007 (ref. 8). *Categories are not mutually exclusive. Cancer Epidemiology, Biomarkers & Prevention 865 Cancer Epidemiol Biomarkers Prev 2007;16(5). May 2007 on April 5, 2021. © 2007 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from http://cebp.aacrjournals.org/ 2007;16:864-865. Cancer Epidemiol Biomarkers Prev Timothy R. Rebbeck, Muin J. Khoury and John D. Potter from Here? Genetic Association Studies of Cancer: Where Do We Go Updated version http://cebp.aacrjournals.org/content/16/5/864 Access the most recent version of this article at: Cited articles http://cebp.aacrjournals.org/content/16/5/864.full#ref-list-1 This article cites 19 articles, 3 of which you can access for free at: E-mail alerts related to this article or journal.Sign up to receive free email-alerts Subscriptions Reprints and .pubs@aacr.orgDepartment at To order reprints of this article or to subscribe to the journal, contact the AACR Publications Permissions Rightslink site. (CCC) Click on "Request Permissions" which will take you to the Copyright Clearance Center's .http://cebp.aacrjournals.org/content/16/5/864 To request permission to re-use all or part of this article, use this link on April 5, 2021. © 2007 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from http://cebp.aacrjournals.org/content/16/5/864 http://cebp.aacrjournals.org/content/16/5/864.full#ref-list-1 http://cebp.aacrjournals.org/cgi/alerts mailto:pubs@aacr.org http://cebp.aacrjournals.org/content/16/5/864 http://cebp.aacrjournals.org/