Multi-Trait Analysis of GWAS and Biological Insights Into Cognition: A Response to Hill (2018)

Abstract

Hill (Twin Research and Human Genetics, Vol. 21, 2018, 84-88) presented a critique of our recently published paper in Cell Reports entitled 'Large-Scale Cognitive GWAS Meta-Analysis Reveals Tissue-Specific Neural Expression and Potential Nootropic Drug Targets' (Lam et al., Cell Reports, Vol. 21, 2017, 2597-2613). Specifically, Hill offered several interrelated comments suggesting potential problems with our use of a new analytic method called Multi-Trait Analysis of GWAS (MTAG) (Turley et al., Nature Genetics, Vol. 50, 2018, 229-237). In this brief article, we respond to each of these concerns. Using empirical data, we conclude that our MTAG results do not suffer from 'inflation in the FDR [false discovery rate]', as suggested by Hill (Twin Research and Human Genetics, Vol. 21, 2018, 84-88), and are not 'more relevant to the genetic contributions to education than they are to the genetic contributions to intelligence'.

A Further Comment on ‘Large-Scale Cognitive GWAS Meta-Analysis Reveals Tissue-Specific Neural Expression and Potential Nootropic Drug Targets’ by Lam et al.

“…Lam et al (2018) assert that their use of multi-trait analysis of genome-wide association study (GWAS) (MTAG; Turley et al, 2018) resulted in the generation of associations that were specific to the trait of cognitive ability. They offer four pieces of supporting evidence based on the genetic correlations derived using linkage disequilibrium score (LDSC) regression (Bulik-Sullivan et al, 2015) found in Table S14 of Lam et al (2018).…”

“…First, Lam et al (2018) state that ‘our Table S14 demonstrates the set of genetic correlations for the MTAG data differ from those of educational attainment (from the Okbay et al (2016) data set) in many cases by an absolute value (for r g ) of 0.10 or greater’. This characterization of the genetic correlations in Table S14 of Lam et al (2017) differing between the MTAG phenotype and education is misleading.…”

“…However, Hill (2018) states in his Table 1, Figure 1, and in the publically available scripts used by Hill (2018) to perform his analyses that genetic correlations of greater than 1 are being treated as 1. In addition, the same reference (Walters, 2016) is used in the rebuttal by Lam et al (2018) to state this is a known issue of LDSC regression, as is used by Hill (2018) to justify the decision to report genetic correlations of greater than 1 as 1. The relevant text from Walters (2016) on how to interpret genetic correlations of greater than 1 is included verbatim here: ‘Thus, far our interpretation has been that as long as artifacts or instability have been ruled out then this result indicates r g approx.…”

“…Figure 1 of Hill (2018) also shows that there was a significant difference between the genetic correlations derived using the MTAG phenotype and those derived using cognitive ability, showing that the genetic architecture of these two traits differ significantly with how they overlap with four measures of education (College completion p = 2.29 × 10 −7 , Years of schooling p = 8.14 × 10 −8 , Years of schooling 2013 p = 9.15 × 10 −8 , Years of schooling 2016 p = 2.70 × 10 −25 ). Neither the fact that the MTAG phenotype shows a genetic correlation near 1 with three measures of education (just the genetic correlations between education and different measures of education were contested) nor the significantly different genetic correlations between the MTAG phenotype and cognitive ability are disputed in the rebuttal by Lam et al (2018). Again, then, it appears that the Lam et al (2017) MTAG phenotype is more similar to education than to cognitive ability.…”

“…Fourth, Lam et al (2018) also believe that the results of their genetic correlations with schizophrenia and bipolar disorder highlight a biologically meaningful set of relationships rather than being the result of their MTAG phenotype having a genetic correlation at unity with education as shown in Figure 1 of Hill (2018). Lam et al (2018) also state that the ‘overall pattern of genetic correlations’ between the three cognitive phenotypes used in Lam et al (2017) — education, the MTAG-intelligence phenotype, and a GWAS composed solely of tests of cognitive ability — is ‘highly similar’.…”