J Abnorm Child Psychol (2013) 41:511–514 DOI 10.1007/s10802-013-9741-0 Editorial Policy for Candidate Gene Studies Charlotte Johnston . Benjamin B. Lahey . Walter Matthys Published online: 20 April 2013 © Springer Science+Business Media New York 2013 Preamble Charlotte Johnston, Editor in Chief and clear policy presented below. I am confident that this policy will serve our journal well as we move forward in The Journal of Abnormal Child Psychology receives an learning about genetic factors in youth psychopathology. increasing number of manuscripts reporting tests of associa- tions of candidate genetic polymorphisms with measures of psychopathology in children and adolescents, including tests Editorial Policy Benjamin B. Lahey and Walter Matthys of gene-environment interactions. These studies represent the exciting new directions that research in child psychopathol- We appreciate the opportunity to articulate an editorial pol- ogy is taking. However, the studies also represent a challenge icy regarding candidate gene studies for the Journal of with regard to the criteria used for evaluating them. Across Abnormal Child Psychology. Although this policy outlines several fields, standards for the publication of such studies are standards for evaluating submitted work, these standards currently in flux. Therefore, it seemed appropriate to advance may be of the greatest value when planning new studies of an editorial policy for the Journal in order to guide both such topics. This policy is largely based on the similar pol- authors and reviewers as we evaluate and incorporate candi- icy adopted by the journal Behavior Genetics (Hewitt 2012), date gene work into the study of child psychopathology. but it is expanded to be directly relevant to this journal. To meet this mission, I invited Dr. Benjamin Lahey a pre- Finally, we note that the emerging nature of candidate gene vious president of ISRCAP, a long-standing editorial board studies will likely necessitate ongoing revision to this policy member, and an active contributor to the research on candi- as methods and techniques advance. date gene and gene-environment interactions in child psy- Much has been written recently about the general issue of chopathology to partner with Dr. Walter Matthys, a current replicability in psychology, psychiatry, and medicine. Serious Associate Editor for the Journal, to prepare a statement out- concerns have been raised about an excessive number of lining such an editorial policy. I express my appreciation to false-positive findings (i.e., findings that have failed to be rep- both of them for their efforts in producing the well-reasoned licated) (Ioannidis 2005c). The concern is that such false pos- itives mislead both scientists and the public, and perhaps C. Johnston (*) direct the allocation of scientific resources in less than opti- Department of Psychology, University of British Columbia, Vancouver, Canada mal ways. There will always be false positives in science, e-mail: [email protected] even when the strongest methods are used, but a number of practices that are commonly engaged in by researchers may B. B. Lahey increase the number of findings that cannot be replicated Departments of Health Studies and Psychiatry and Behavioral Neuroscience, University of Chicago, Chicago, IL, USA (Simmons et al. 2011). Moreover, there is good reason to e-mail: [email protected] believe that science is not quick to detect false-positives and correct misimpressions (Ioannidis 2012). Thus, the issue of W. Matthys replicability is front and center today in many sciences. Department of Child and Adolescent Studies, Utrecht University, Utrecht, The Netherlands The tendency for novel findings to subsequently fail rep- e-mail: [email protected] lication may be particularly great in new and “hot” areas of 512 J Abnorm Child Psychol (2013) 41:511–514 research, such as candidate gene associations and gene- each candidate genetic variant represents only one of an environment interactions. The existence of a strong publica- extremely large number of possible variants, a strong tion bias towards positive findings is partly due to incentives case must be made for studying the selected variant and both for authors and editors to publish positive reports. its associated psychological function over other possible Other things being equal, reviewers and editors may be choices. more likely to agree that exciting and novel findings should 3. Attention to population heterogeneity. A satisfactory be published than research on more established topics. approach to population heterogeneity is required. Molecular Particular attention has been given to the problem of rep- genetic studies pose special problems for sample selec- licability in research using molecular genetic methods. tion. On the one hand, there are important advantages Although there are many ways to study associations between inherent in studying diverse samples, if only to conduct genetic variants and behavior, each with its own pitfalls, research that applies to all children and adolescents. On particular concerns have been raised about replicability in the other hand, the world’s ancestry groups differ in candidate gene studies (Ioannidis 2005a, 2005b; Ioannidis terms of both allele frequencies and complexity. This and Khoury 2011; Ioannidis et al. 2001; Munafo 2009; creates opportunities for mistaken findings based on Sullivan 2007). This is not to say that there have not been population stratification (Rosenberg et al. 2010) and may replicated candidate gene findings; there have been (Gizer even require different genotyping platforms for different et al. 2009). Nonetheless, the number of statistically signifi- ancestry groups (Hoffmann et al. 2011). These issues are cant candidate gene associations and gene-environment particularly problematic for smaller diverse samples that interactions that appear not to be replicated is easily large rarely have the power to successfully address these enough to warrant particular concern (Duncan and Keller issues. 2011; Hewitt 2012; Hunter 2005; Mill and Petronis 2007; 4. Prospective designs. Prospective longitudinal designs Moffitt et al. 2005). provide much stronger tests of genetic associations Keeping in mind that any set of restrictive standards and gene-environment interaction than cross-sectional designed to reduce the number of published false positives designs. This is because prospective studies can rule out will almost certainly have the disadvantage of reducing the reverse causation (Kraemer 2010). It may be possible to number of published novel findings that are later replicated, justify cross-sectional studies of genetic associations and the position of the Journal is that special care needs to be gene-environment interaction, however. As noted below, taken in making editorial decisions in a manner that reduces these must fully take gene-environment correlation into false-positive findings of candidate gene associations and account, which is difficult to accomplish in cross- sectional gene-environment interactions. Therefore, manuscripts involv- designs. ing candidate genes will only be favorably viewed if they 5. Statistical power and control of alpha. Manuscripts meet the standards below. These standards are not entirely will be considered for publication only if the tests are new to the journal, but rather reflect a synthesis of the stan- based on sample sizes that provide sufficient statistical dards used by reviewers and editors in recent editorial deci- power based on reasonable estimates of effect sizes. The sions. Sound arguments for special cases will, of course, be sample size must allow statistical analyses to correct the considered, but the following list is intended as appropriate alpha level for the number of related statistical tests per- standards. formed, even ones that are not reported (Ioannidis 2005c; Little et al. 2009). 1. Measurement. As in any area of research, tests of 6. Statistical tests of gene-environment interaction. genetic associations and gene-environment interactions Special care needs to be taken when conducting any sta- require reliable and valid measures of all variables. tistical test of interaction (McClelland and Judd 1993). Findings of gene-environment interactions are strength- Because violations of the assumption of multivariate ened, moreover, when the environments and phenotypes normality can result in the “detection” of interactions are measured using separate informants (Moffitt et al. when none exists (Eaves 2006), statistical methods must 2005). reduce the likelihood of scaling artifacts. This is espe- 2. Biological and psychological plausibility. In studies of cially pertinent for psychopathology research because candidate genetic variants and gene-environment inter- highly skewed and kurtotic distributions of dimensions actions, findings are more convincing when the genetic of psychopathology almost never meet the assumption variants, environments, and phenotypes are all selected of multivariate normality. In addition, tests of gene- based on a well articulated and plausible biological and environment can be difficult to interpret in the presence psychological model of the phenomenon. Choosing of gene-environment correlation (Eaves et al. 2003; environments and genetic variants with known func- Rathouz et al. 2008). Thus, tests of interaction must tional consequences strengthens such a case. Because take this phenomenon into account. Gene-environment J Abnorm Child Psychol (2013) 41:511–514