It's a good time to be a behavioral geneticist - twin study estimates of complex traits' heritability vindicated yet again; "Substantial SNP-based heritability estimates for working memory performance", Vogler et al 2014:
"Working memory (WM) is an important endophenotype in neuropsychiatric research and its use in genetic association studies is thought to be a promising approach to increase our understanding of psychiatric disease. As for any genetically complex trait, demonstration of sufficient heritability within the specific study context is a prerequisite for conducting genetic studies of that trait. Recently developed methods allow estimating trait heritability using sets of common genetic markers from genome-wide association study (GWAS) data in samples of unrelated individuals. Here we present single-nucleotide polymorphism (SNP)-based heritability estimates (h2SNP) for a WM phenotype. A Caucasian sample comprising a total of N=2298 healthy and young individuals was subjected to an N-back WM task. We calculated the genetic relationship between all individuals on the basis of genome-wide SNP data and performed restricted maximum likelihood analyses for variance component estimation to derive the h2SNP estimates. Heritability estimates for three 2-back derived WM performance measures based on all autosomal chromosomes ranged between 31 and 41%, indicating a substantial SNP-based heritability for WM traits. These results indicate that common genetic factors account for a prominent part of the phenotypic variation in WM performance. Hence, the application of GWAS on WM phenotypes is a valid method to identify the molecular underpinnings of WM.
...Working memory (WM) has a pivotal role in human cognition and cognitive performance, allowing the integration of information from instantly perceived stimuli, long-term-memory and thought-processes. A wide variety of psychological testing procedures have been developed to render WM performance quantifiable.1, 2, 3, 4 Assessment of WM performance along with brain imaging techniques have been used to elucidate and delimit its different components, aiming at a more circumscribed understanding of the neuronal processes at the basis of WM.5 The integration of extensive research on WM yields a common consensus that it constitutes a complex trait and that the buffer size for the transiently stored content varies interindividually, which is partly due to genetic factors. WM is an important endophenotype in neuropsychiatric research and its use in genetic association studies is thought to be a promising approach to increase our understanding of psychiatric disease.6, 7, 8, 9, 10, 11 Namely, two recent studies have corroborated the genetic link between schizophrenia and WM, demonstrating the validity of this endophenotype for schizophrenia research: Stefansson et al.12 report that control subjects carrying Copy Number Variants, which predispose to schizophrenia and autism, perform at a level that is in between patients and population controls in cognitive tasks including a spatial WM test. A genome-wide gene set enrichment study conducted by our group, identified a set of voltage-gated cation channel activity genes that were robustly linked to performance in WM tasks in healthy individuals and also to risk for schizophrenia in a large case–control sample.13 Both these studies suggest that the findings of cognitive deficits are translatable between healthy subjects and cohorts of psychiatric patients. Improving our understanding of the molecular basis of this endophenotype might be key for future drug discovery and treatment options in psychiatry.14,15 As for any genetically complex trait, demonstration of sufficient heritability within the specific study context is a prerequisite for conducting genetic studies of that trait. Heritability is a concept that summarizes how much of the phenotypic variation in a given trait is attributable to heritable factors, the majority of them being genetic. Since inter-individual trait differences attributable to genetic variability are a prerequisite for quantitative trait loci mapping, estimation of trait heritability is important to demonstrate the validity of a quantitative trait loci study of a given trait. Conventionally, heritability estimates in humans are derived from phenotypic data by comparing correlations between relatives, where the extent of genetic relatedness is derived from the degree of relationship. Results from previous twin and family studies that used a variety of tasks to measure WM performance have shown heritability estimates ranging between 15 and 72%.16, 17, 18, 19 Recently developed methods propose inferring genetic identity from high-throughput SNP data and correlating these estimates with phenotypic resemblance among unrelated individuals.20,21 This allows the estimation of an SNP-based heritability measure (h2SNP) for any specific genome-wide association study data set. Importantly, the h2SNP value quantifies the amount of phenotypic variation that can be explained by the common SNPs represented in genome-wide association study data sets. Of note, h2SNP hereby represents a lower bound heritability estimate, because causal variants that are neither genotyped nor tagged by the used set of markers are completely disregarded. In the present study we show that a substantial part of phenotypic variance of WM performance can be explained by using the common marker set represented on the Affymetrix 6.0 Human SNP array to estimate h2SNP values for N-back-derived WM phenotypes.
...The h2SNP estimates for the N-back-derived WM phenotypes range from 31 to 41% with a mean standard error of 0.14. Thus, the h2SNP estimates are consistent with the previously reported heritability for WM on the basis of twin-studies.27 Taken together, these findings add further support for the hypothesis that WM is a highly heritable trait. The chromosomal partitioning analysis (Figure 2), which depicts h2SNP estimates for all chromosomes analyzed individually, provides a clear hint for the pronounced polygenicity of the investigated WM phenotype. This finding is in line with the previously reported ubiquitious polygenicity of human complex traits.28...Of note, the heritability estimation for complex traits using genome-wide data sets is rather a complement than a substitution to studies on twin- and family-based heritability. Marker-based heritability estimation represents a lower bound for the true trait heritability as it relies only on the effects of common variants assuming an additive variance model. On one hand, the investigation of large pools of unrelated individuals allows to assess heritability without the undermining effects due to shared environment or familiality. Yet, on the other hand, it will not take the potential effects of rare variants into account. The SNP-based heritability estimates for WM suggest that a large share of phenotypic variance can be explained by common SNPs rendering well-powered genome-wide association study data sets a promising tool to discover molecular players that act in concert to form this complex trait."
Note, incidentally, that because WM correlates highly with IQ, their dataset could be used to boost a GWAS for IQ variants in the same fashion as Rietveld et al 2014 (https://plus.google.com/103530621949492999968/posts/ZdiijNxg78v) although given it's only 2.7k genotypes, probably not worth anyone's while to navigate the consent issues.
"Working memory (WM) is an important endophenotype in neuropsychiatric research and its use in genetic association studies is thought to be a promising approach to increase our understanding of psychiatric disease. As for any genetically complex trait, demonstration of sufficient heritability within the specific study context is a prerequisite for conducting genetic studies of that trait. Recently developed methods allow estimating trait heritability using sets of common genetic markers from genome-wide association study (GWAS) data in samples of unrelated individuals. Here we present single-nucleotide polymorphism (SNP)-based heritability estimates (h2SNP) for a WM phenotype. A Caucasian sample comprising a total of N=2298 healthy and young individuals was subjected to an N-back WM task. We calculated the genetic relationship between all individuals on the basis of genome-wide SNP data and performed restricted maximum likelihood analyses for variance component estimation to derive the h2SNP estimates. Heritability estimates for three 2-back derived WM performance measures based on all autosomal chromosomes ranged between 31 and 41%, indicating a substantial SNP-based heritability for WM traits. These results indicate that common genetic factors account for a prominent part of the phenotypic variation in WM performance. Hence, the application of GWAS on WM phenotypes is a valid method to identify the molecular underpinnings of WM.
...Working memory (WM) has a pivotal role in human cognition and cognitive performance, allowing the integration of information from instantly perceived stimuli, long-term-memory and thought-processes. A wide variety of psychological testing procedures have been developed to render WM performance quantifiable.1, 2, 3, 4 Assessment of WM performance along with brain imaging techniques have been used to elucidate and delimit its different components, aiming at a more circumscribed understanding of the neuronal processes at the basis of WM.5 The integration of extensive research on WM yields a common consensus that it constitutes a complex trait and that the buffer size for the transiently stored content varies interindividually, which is partly due to genetic factors. WM is an important endophenotype in neuropsychiatric research and its use in genetic association studies is thought to be a promising approach to increase our understanding of psychiatric disease.6, 7, 8, 9, 10, 11 Namely, two recent studies have corroborated the genetic link between schizophrenia and WM, demonstrating the validity of this endophenotype for schizophrenia research: Stefansson et al.12 report that control subjects carrying Copy Number Variants, which predispose to schizophrenia and autism, perform at a level that is in between patients and population controls in cognitive tasks including a spatial WM test. A genome-wide gene set enrichment study conducted by our group, identified a set of voltage-gated cation channel activity genes that were robustly linked to performance in WM tasks in healthy individuals and also to risk for schizophrenia in a large case–control sample.13 Both these studies suggest that the findings of cognitive deficits are translatable between healthy subjects and cohorts of psychiatric patients. Improving our understanding of the molecular basis of this endophenotype might be key for future drug discovery and treatment options in psychiatry.14,15 As for any genetically complex trait, demonstration of sufficient heritability within the specific study context is a prerequisite for conducting genetic studies of that trait. Heritability is a concept that summarizes how much of the phenotypic variation in a given trait is attributable to heritable factors, the majority of them being genetic. Since inter-individual trait differences attributable to genetic variability are a prerequisite for quantitative trait loci mapping, estimation of trait heritability is important to demonstrate the validity of a quantitative trait loci study of a given trait. Conventionally, heritability estimates in humans are derived from phenotypic data by comparing correlations between relatives, where the extent of genetic relatedness is derived from the degree of relationship. Results from previous twin and family studies that used a variety of tasks to measure WM performance have shown heritability estimates ranging between 15 and 72%.16, 17, 18, 19 Recently developed methods propose inferring genetic identity from high-throughput SNP data and correlating these estimates with phenotypic resemblance among unrelated individuals.20,21 This allows the estimation of an SNP-based heritability measure (h2SNP) for any specific genome-wide association study data set. Importantly, the h2SNP value quantifies the amount of phenotypic variation that can be explained by the common SNPs represented in genome-wide association study data sets. Of note, h2SNP hereby represents a lower bound heritability estimate, because causal variants that are neither genotyped nor tagged by the used set of markers are completely disregarded. In the present study we show that a substantial part of phenotypic variance of WM performance can be explained by using the common marker set represented on the Affymetrix 6.0 Human SNP array to estimate h2SNP values for N-back-derived WM phenotypes.
...The h2SNP estimates for the N-back-derived WM phenotypes range from 31 to 41% with a mean standard error of 0.14. Thus, the h2SNP estimates are consistent with the previously reported heritability for WM on the basis of twin-studies.27 Taken together, these findings add further support for the hypothesis that WM is a highly heritable trait. The chromosomal partitioning analysis (Figure 2), which depicts h2SNP estimates for all chromosomes analyzed individually, provides a clear hint for the pronounced polygenicity of the investigated WM phenotype. This finding is in line with the previously reported ubiquitious polygenicity of human complex traits.28...Of note, the heritability estimation for complex traits using genome-wide data sets is rather a complement than a substitution to studies on twin- and family-based heritability. Marker-based heritability estimation represents a lower bound for the true trait heritability as it relies only on the effects of common variants assuming an additive variance model. On one hand, the investigation of large pools of unrelated individuals allows to assess heritability without the undermining effects due to shared environment or familiality. Yet, on the other hand, it will not take the potential effects of rare variants into account. The SNP-based heritability estimates for WM suggest that a large share of phenotypic variance can be explained by common SNPs rendering well-powered genome-wide association study data sets a promising tool to discover molecular players that act in concert to form this complex trait."
Note, incidentally, that because WM correlates highly with IQ, their dataset could be used to boost a GWAS for IQ variants in the same fashion as Rietveld et al 2014 (https://plus.google.com/103530621949492999968/posts/ZdiijNxg78v) although given it's only 2.7k genotypes, probably not worth anyone's while to navigate the consent issues.