Pleiotropy between language impairment and broader behavioral disorders—an investigation of both common and rare genetic variants

Table 1 Results of the PRS regression analyses in iPSYCH; updated results from the pilot study sample are shown for comparison

Phenotype	R²/Nagelkerke’s R² (%)	Adjusted R² (%)	Scaled PRS beta/ odds ratio (95% CI)	P value
iPSYCH sample^*
ASD	0.017	0.007	1.02 (1.00–1.05)	0.037
Childhood autism	0.079	0.042	1.06 (1.02–1.10)	0.001
Asperger’s syndrome	0.002	0.001	1.01 (0.98–1.04)	0.552
ADHD	0.002	0.002	1.01 (0.99–1.03)	0.426
Schizophrenia	0.002	0.001	0.99 (0.95–1.03)	0.613
Pilot study sample^**
ASD	0.017	0.017	0.97 (0.59–1.57)	0.887
ADHD	0.008	0.009	1.02 (0.75–1.39)	0.897
SLI	3.5	5.09	1.60 (1.02–2.54)	0.041
Height	0.006	NA	0.07 ((− 0.98) –1.13)	0.895

The adjusted R² includes an adjustment for the prevalence and proportion of cases in each sample, as implemented in the PRSice software. Associations in boldface survive Bonferroni correction for multiple testing (not applicable to positive or negative controls). Associations in italic are only nominally significant (not applicable to positive or negative controls)
PRS polygenic risk score, CI confidence interval, ASD autism spectrum disorder, ADHD attention deficit/hyperactivity disorder, SLI specific language impairment, NA not applicable
The R² reported for height is a true R² from a linear regression and not a Nagelkerke’s R², and the effect is the beta (regression coefficient) and not an odds ratio; it should therefore be judged relative to 0 and not to 1. The height variable was collected in different ways at the Copenhagen and Aarhus centers for the VIA 7 study; the Aarhus center used, for the most part, a uniform method of measuring height with a special device, but we note that the PRS is not significantly predictive of height even if we restrict the analysis to include only children measured in Aarhus. We include this for the sake of completeness, even though the sample size is much smaller and thus the estimates are less reliable (R² = 0.7%, P = 0.419)
*72,932 SNPs were included in the PRS
**80,121 SNPs were included in the PRS

ISSN: 1866-1955