“Forward Genetics” as a Method to Maximize Power and Cost-Efficiency in Studies of Human Complex Traits
- First Online:
- Cite this article as:
- Boks, M.P.M., Derks, E.M., Dolan, C.V. et al. Behav Genet (2010) 40: 564. doi:10.1007/s10519-010-9348-y
- 306 Downloads
There is increasing interest in methods to disentangle the relationship between genotype and (endo)phenotypes in human complex traits. We present a population-based method of increasing the power and cost-efficiency of studies by selecting random individuals with a particular genotype and then assessing the accompanying quantitative phenotypes. Using statistical derivations, power- and cost graphs we show that such a “forward genetics” approach can lead to a marked reduction in sample size and costs. This approach is particularly apt for implementing in epidemiological studies for which DNA is already available but the phenotyping costs are high.
KeywordsPower Forward genetics Complex traits Selective sampling Phenomics
Several genome-wide association studies have recently identified novel susceptibility loci for medical conditions such as diabetes mellitus and schizophrenia (Barret et al. 2009; Stefansson et al. 2009). This has increased the need to investigate the phenotypic differences that are conferred by such quantitative trait loci. However, due to the small to modest contributions of single loci to most complex traits, such phenotypic differences are hard to detect. The problem of small effect sizes of specific alleles or haplotypes is compounded by the complexity of many of the traits of interest. Most candidate genes have been identified in subjects with a disorder that incorporates a broad variety of symptoms. The co-occurrence of several symptoms at once can be the result of pleiotropic effects of a singe variant, but may also be due to the underlying abnormality. Many disorders coincide with physical, emotional, and social abnormalities; for example, depression is associated with cognitive problems, cardiovascular risks and social problems, among others. These concomitant phenomena are likely to influence the expression of the original trait, and may well obscure the initial relationship between a candidate gene and a trait. As a consequence, the particular symptoms or abnormalities associated with these genes remain unclear.
We, therefore, propose reversing the process: instead of selecting a trait and examining its relationship with the underlying genes, we will select genetic variation and examine the accompanying trait. Testing the influence of a particular gene on phenotypes is a common approach in both animal and molecular research, where the influence of genetic variation is often studied by inbreeding the genetic variant by creating a knockout mouse, or by transposing the variant of interest into cell cultures or organisms by means of a vector. The statistical power of an association test for a candidate gene depends on the distribution of genotypes in the test population. A maximum statistical power for a given number of phenotyped individuals is obtained when the test population consists of equal numbers of alternative homozygotes at the candidate gene (as is the case when the two alleles at the candidate gene are of equal frequency). However, since allele frequencies at the candidate gene locus are generally far from equal, the distribution of informative alleles in the population as a whole is generally far from this optimal distribution. Thus, depending on the relative costs of determining phenotypes compared to the cost of genotypes, it may be more effective to genotype a large sample population and then choose a set of individuals with an optimal distribution of genotypes for further phenotyping. Here we provide information on the statistical power under different genotype sampling strategies, as a function of explained variance, dominance and allele frequency at the candidate gene, and on phenotype/genotype cost ratios.
Selecting subjects from the general population based on homozygosity for a candidate gene instead of subjects with an apparent disorder has two major advantages. (1) It means the investigation of the relationship with genotype is unbiased by selection for severity of disease, and we therefore avoid bias as a result of secondary symptoms. (2) This approach facilitates the estimation of the effects of single variants in relative isolation, because the selection is not based on phenotype. As a consequence, there is no selection for the presence of additional risk variants for that particular phenotype although it will shift the distribution of the phenotype.
The value of such a “forward genetics” approach is seen in the increase in statistical power and its cost-effectiveness. As already pointed out, the increase in power is due to the selection of the most informative subjects. Other strategies, such as the extreme discordant and concordant design (Risch and Zhang 1995), in essence do the same by selecting extreme phenotypes. With the ever reducing costs of genotyping, our strategy only has merit if the cost of obtaining phenotype information is high. In studies of complex and quantitative phenotypes, such as those that apply costly neuroimaging, this approach can be particularly advantageous.
We investigated the sample size requirement of this approach and the cost-effectiveness under different scenarios.
Finally, we calculated the non-centrality parameter (NCP), i.e., F*(N − k), and the statistical power based on this parameter and a chosen alpha. We compared the NCP calculated with that provided by the genetic power calculator (Purcell et al. 2003). The application can be extended for scenarios where more genes are considered (see supplementary R script). We performed illustrative power and costs analyses under several scenarios, including one in which we selected three genes and subjects were available that were homozygous for two of the rare homozygous variants and heterozygous for the third variant compared to subjects that were homozygous for all three common variants. The calculations were carried out in R (R Development Core Team 2005).
We have demonstrated that “forward genetics” is a cost-effective method for investigating genotype–phenotype associations in complex human traits. The figures, formulas and R-script provided can be used to estimate the sample size required under different conditions and our calculations show that a marked increase in power and reduction of costs can be obtained by this approach, particularly for those scenarios with relatively high phenotyping costs. Considering the ever reducing costs of genotyping, and the use of increasingly complex phenotypes, such as measures of brain volume obtained with MRI, these scenarios are frequently encountered. If, for instance, after pooling of data on hippocampal volume in a consortium (covering several hundreds of subjects), genome-wide analysis yields a few, significantly associated loci, it would make sense to type these loci in a large population sample and to obtain MRI scans in two groups based on the selection of all these SNPs in order to maximize power. Clearly, such a replication sample could also be used for further genome-wide analysis, but at least the sample would be optimal for replicating the effects found. This method, therefore, is a cost-effective way to confirm previously reported genetic association signals, to refine genotype–phenotype relationships, and thus to facilitate the discovery of susceptibility loci in complex quantitative traits. Extensions of this design, in which subjects are selected on the basis of two or more genetic variants, have been presented here. Interaction between the genes can subsequently be tested as a measure of epistasis.
Applying forward genetics can be particularly helpful in studies where DNA is available as part of an epidemiological (longitudinal) study. We plan to use this approach for the study of cognitive and brain volumetric phenotypes in a large, population-based epidemiological sample from the Netherlands. This study will be performed as part of the Utrecht Health Project (Grobbee et al. 2005). In this design DNA of more than 6,000 subjects is available for genotyping. Subjects homozygous for alleles that are associated with cognitive traits, and subjects homozygous for the wildtype allele, will be invited for further comprehensive assessments. Our power analyses show that a number of N = 90 per genetic variant will be sufficient to obtain a type I error rate of 0.007 and a power of 0.8. We plan to comprehensively assess the psychological-, cognitive-, and social functioning of the selected subjects in order to test seven different phenotypes. The raters and participants will remain blind to the subject’s genotype status in order to avoid phenotyping bias and ethical complications. The classical approach would be to assess a large numbers of participants from the epidemiological study and to perform a case–control association analysis or quantitative trait locus (QTL) analysis. As pointed out above, the costs of assessing the large numbers of participants required in the conventional design would be much higher compared to the forward genetics approach. Particularly in cases where the genetic variant is rare and many assessments would be required to obtain sufficient power.
Our approach differs from the recently advocated “reverse phenotyping” approach (Schulze and McMahon 2004), in which both genotype and phenotype information is assumed to be available. The phenotype–genotype relationship is, however, re-assessed by revisiting phenotypes (reverse phenotyping). Clearly, a strong argument for Schulze and McMahon’s approach (Schulze and McMahon 2004) can be made and there are several recent studies that have successfully applied such reverse phenotyping (Tilley et al. 1998; Silverman et al. 2002; Schaid et al. 2006). However reverse phenotyping cannot solve the power issue, which is relevant to the investigation of rarer genetic variants; nor does it result in a cost reduction in the study of expensive phenotypes, such as those based on brain imaging. Overall, our analyses demonstrate that forward genetics can be a useful tool for studying more complex and expensive behavioral phenotypes in the context of epidemiological studies.
We are grateful to J.L. Senior for editing the manuscript.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.