Paper III

Shared genetic architecture between mental health and the brain functional connectome in the UK Biobank

After deriving profiles of mental health and uncovering their genetic architecture in the first paper, and delineating the genetic architecture of the brain functional connectome in the second, the last paper aimed to connect the two by investigating the genetic link between the brain functional connectome and the mental health-related symptom profiles. We performed multivariate GWAS analysis on the component loadings of independent components from the Roelfs et al.¹ paper using MOSTest. This resulted in a single measure capturing different facets relevant for mental health. These analyses identified 10 significant genetic loci associated with this feature which were linked through functional mapping to brain structures such as the cerebellum and the amygdala. When we analyzed whether the loci in each of the psychiatric disorder GWAS were also nominally significant (p < 0.05) in the multivariate measures of mental health we found that of the 67 loci in the psychiatric disorders, 63 were nominally significant in the multivariate measure of mental health profiles.

Comparing the geneset from this multivariate measure of mental health with the genesets from a number of psychiatric disorders we found 9 overlapping genes, 7 for BIP, 1 for ADHD, and 1 for SCZ. Using conjFDR we identified 35 shared loci between the psychiatric disorders and the multivariate measure of mental health, 10 for BIP, 8 for both MD and ADHD, 5 for SCZ, and 4 for ASD.

Subsequently, we also calculated the genetic overlap between the multivariate measure of mental health and the two multivariate brain connectivity measures generated in the Roelfs et al.² paper. This analysis yielded 18 shared loci between the mental health profiles with FC and 5 shared loci with node variance as illustrated in Figure 27.1. Functional mapping of the genes associated with these shared loci revealed pathways involved in a number of processes relevant for neurobiology, such as axonal growth regulation (ngfr and rhoa) and regulation of certain transcription factors (mef2c).

Figure 27.1: Manhattan plots showing the shared genetic architecture derived using conjFDR between the multivariate definition of the brain functional connectome and a multivariate measure of mental health. Originally published in Roelfs et al.³.

In summary, we identified shared genetic architecture between independent symptom profiles reflecting different aspects of mental health, between a multivariate measure of the functional connectome and psychiatric disorders, and between the symptom profiles and the brain connectome. These discoveries taken together highlight the scientific value that can be derived from multimodel approaches in population-level studies.

1.

Roelfs D, Alnæs D, Frei O, van der Meer D, Smeland OB, Andreassen OA, et al. Phenotypically independent profiles relevant to mental health are genetically correlated. Translational Psychiatry [Internet]. 2021 Apr 1;11(1):202. Available from: https://doi.org/10.1038/s41398-021-01313-x

2.

Roelfs D, van der Meer D, Alnæs D, Frei O, Shadrin AA, Loughnan R, et al. Genetic overlap between multivariate measures of human functional brain connectivity and psychiatric disorders. Nature Mental Health [Internet]. 2024 Feb 1;2(2):189–99. Available from: https://doi.org/10.1038/s44220-023-00190-1

3.

Roelfs D, Frei O, van der Meer D, Tissink E, Shadrin A, Alnaes D, et al. Shared genetic architecture between mental health and the brain functional connectome in the UK Biobank. BMC Psychiatry [Internet]. 2023 Jun 23;23(1):461. Available from: https://doi.org/10.1186/s12888-023-04905-7