Omics, Biomarkers, and Aggressive Behavior

Research output: PhD ThesisPhD-Thesis - Research and graduation internal

337 Downloads (Pure)

Abstract

To investigate the biological etiology of human aggressive behavior through biomarker discovery and to explore the genetic etiology of variation in blood metabolite levels in humans, I analyzed twin-family data collected by the Netherlands Twin Register (NTR) and data of children referred to an academic centrum for child- and adolescent psychiatry. In my thesis, I argue for combining the classical twin design with genome-wide SNP (Single-Nucleotide Polymorphism) data to simultaneously estimate the total and SNP-based heritability. When analyzing 361 blood metabolites measured in a cohort of 5117 twin-family members, I applied this approach and found a stronger contribution of metabolite loci to lipids than to organic acids, and showed that the variance explained by metabolite loci differs significantly among metabolite classes and lipid species. These approaches based on measured SNPs are unsuitable for investigating the contribution of shared environmental factors. Therefore, I applied the classic twin design to 237 blood metabolite levels in 886 monozygotic (MZ) and 601 dizygotic (DZ) adult twin pairs, and reported significant estimates of the shared environment for only 6 out of 237 metabolites. To conclude, additive genetic factors contribute to approximately 50% of the individual variance in blood metabolites levels in adults. To look at aggression and biochemical biomarkers, I began with reviewing the existing literature for human aggression and related traits and found associations with biomarkers of multiple classes, including inflammation markers, neurotransmitters, lipoproteins, and thyroid and steroid hormones. However, I did not replicate the associations of lipid, inflammation, and related biomarkers with adult aggression in a twin-family cohort (N=5588), but found differences in glucose, fibrinogen, C-reactive protein, interleukin-6, and low-density lipoprotein cholesterol levels in the extremely discordant MZ twin pairs (12 pairs). I also tested the hypothesis that lipid CpG sites are differentially methylated in adult aggression, but this hypothesis was not supported. I coordinated the large-scale collection of urine and buccal-cell samples for (epi)genomic and metabolomic measurements in young twin pairs selected on aggression in the NTR, and share the protocols in my thesis. When analyzing these data, I observed associations of the urinary metabolites o-phosphoserine and gamma-L-glutamyl-L-alanine with childhood aggression in the discovery twin cohort (N=783). These findings did not replicate in an independent sample of discordant twin pairs (N=378) or in a cohort of clinical cases and matched twin controls (N=367). The top associated metabolites hint at a role of metabolic dysregulation of neurotransmission, energy metabolism, and oxidative stress in childhood aggression. Next, I performed an integrated multi-omics analysis of childhood aggression based on urinary metabolites, genome-wide DNA methylation, and SNP-based transmitted and non-transmitted polygenic scores for aggression and correlated traits. The metabolite data included a new platform targeting steroid hormones. Training of multi-omics models for childhood aggression showed good in-sample prediction through cross-validation (N=645), but out-of-sample prediction in hold-out (N=277) and clinical samples (N=142) was poor. The high correlations among the omics traits selected in the multi-omics models confirmed associations of childhood aggression with known risk factors and provided novel insight into the correlational structure of omics traits from different omics layers. For example, we showed that indirect genetic effects correlate most strongly with DNA methylation and amino acid levels, suggesting that parental genotypes influence offspring omics traits through the rearing environments parents create for their offspring. In summary, genetic factors influence on blood metabolite levels, and the shared environment contributes marginally to familial resemblance in blood metabolites. For childhood aggression, heritability is around 50%. When combining aggression and biomarker information, we see broad biological pathways that appear to play a role in human aggressive behavior, including dysregulation of neurotransmission, inflammation, and lipid, steroid hormone, and energy metabolism pathways.
Original languageEnglish
QualificationDr.
Awarding Institution
  • Vrije Universiteit Amsterdam
Supervisors/Advisors
  • Boomsma, Dorret, Supervisor
  • Bartels, Meike, Supervisor
  • van Dongen, Jenny, Co-supervisor
Award date14 Mar 2022
Publisher
Print ISBNs9789493270473
Electronic ISBNs9789493270473
Publication statusPublished - 14 Mar 2022

Keywords

  • aggression
  • biomarkers
  • omics
  • metabolomics
  • twins
  • genetics
  • heritability
  • epigenetics
  • polygenic scores
  • multi-omics

Fingerprint

Dive into the research topics of 'Omics, Biomarkers, and Aggressive Behavior'. Together they form a unique fingerprint.

Cite this