Genetic estimates of correlation and causality between blood-based biomarkers and psychiatric disorders

Abstract

There is a long-standing interest in exploring the relationship between blood-based biomarkers of biological exposures and psychiatric disorders, despite their causal role being difficult to resolve in observational studies. In this study, we leverage genome-wide association study data for a large panel of heritable biochemical traits measured from serum to refine our understanding of causal effect in biochemical-psychiatric trait parings. In accordance with expectation we observed widespread evidence of positive and negative genetic correlation between psychiatric disorders and biochemical traits. We then implemented causal inference to distinguish causation from correlation and found strong evidence that C-reactive protein (CRP) exerts a causal effect on psychiatric disorders, along with other putatively causal relationships involving urate and glucose. Strikingly, these analyses suggested CRP has a protective effect on three disorders including anorexia nervosa, obsessive-compulsive disorder, and schizophrenia, whilst being a risk factor for major depressive disorder. Multivariable models that conditioned CRP effects on interleukin-6 signalling and body mass index suggested that CRP-schizophrenia relationship was not likely mediated by those factors. Collectively, these data suggest that there are shared pathways that influence both biochemical traits and psychiatric illness, including factors such as CRP that are likely to constitute a causal effect and could be targets for therapeutic intervention and precision medicine.

Genetics-informed precision treatment formulation in schizophrenia and bipolar disorder

“…While this association does not survive multiple testing correction, and thus should be interpreted cautiously, it is notable as the FADS1 network PES was associated with lipid profiles in an analogous direction to what has previously shown to be genetically correlated with anorexia nervosa GWASs via LD score regression. 11,21 In summary, these data coupled with the biochemical associations support the distinct nature of network PESs from PRSs and emphasize the unique insights that can be afforded by these partitioned scores.…”

“…While we addressed this previously, in respiratory medicine, by triangulating prioritized pathways through causal inference of pharmacologically sensitive biochemical traits, 19 these relationships are more difficult to find in psychiatric disorders. 21 In the current study, we propose a new implementation of the PES that is informed by genetically proxied mRNA or protein expression of drug target genes. This approach identifies candidate psychiatric drug repurposing opportunities at the population level that can then be more appropriately integrated with genetic risk scores relevant to these target genes, to identify individuals who may benefit more readily from these compounds.…”

Directional anchor genes refine polygenic informed treatment selection in schizophrenia and bipolar disorder

“…There were also several other nominal associations ( P < 0.05), including one of particular interest in the case of the BIP FADS1 network PES, for which a higher score displayed some evidence of a protective effect on self-reported anorexia nervosa. Whilst this association does not survive multiple testing correction, and thus should be interpreted cautiously, it is notable as the FADS1 network PES was associated with lipid profiles in an analogous direction to what has previously shown to be genetically correlated with anorexia nervosa GWAS via LD score regression [11, 21]. In summary, these data coupled with the biochemical associations support the distinct nature of network PES from PRS and emphasise the unique insights that can be afforded by these partitioned scores.…”

“…However, a limitation of the PES approach is that it is not innately informative as to which of the suite of drugs targeting a pathway will be most useful, particularly in regards as to whether an agonist or antagonist of target genes should be investigated. While we addressed this previously, in respiratory medicine, by triangulating prioritised pathways through causal inference of pharmacologically sensitive biochemical traits [19, 20], these relationships are more difficult to find in psychiatric disorders [21]. In the current study, we propose a new implementation of the PES that is informed by genetically proxied mRNA or protein expression of drug target genes.…”

Multiomic prioritisation of risk genes for anorexia nervosa

“…Accumulating observational evidence suggests a relationship may exist between immune function and anorexia nervosa (47, 48). Starvation can lead to inflammation; however, there remains differences between the presentation of immune dysregulation during malnutrition and anorexia nervosa, suggesting an underlying relationship may exist beyond the pathology associated with malnutrition (49).…”