Andrew Joseph Schork

Anti-cytokine autoantibodies are increasingly recognized as modulators of immune function and determinants of disease risk, yet their genetic basis and population-level impact remain unclear. Here we perform genome-wide association studies of autoantibodies against IL-1α, IL-6, IL-10, IFN-α, IFN-β, IFN-γ and GM-CSF in 15,000 individuals from the Danish Blood Donor Study. We identify 52 genome-wide significant loci, implicating genes enriched in antigen-presentation pathways. Using these data, we derive cytokine-specific polygenic risk scores and evaluate their associations across 300,000 individuals in the Copenhagen Hospital Biobank. Genetic predisposition to IL-1α autoantibodies is associated with reduced risk of rheumatoid arthritis and certain cancers, whereas genetic predisposition to IL-6 autoantibodies is associated with increased risk of diabetes, mirroring prior clinical observations. These findings define the genetic architecture of anti-cytokine autoantibodies and indicate their divergent effects on human disease risk at population scale, providing a framework for mechanistic investigation and potential clinical stratification.

Studies of resemblance for disorders and other traits measured at the binary (yes/no) level between relatives frequently contain individuals who are currently in the negative category but who will become positive in future. For example, a 10-year-old may develop depression in the future, but is as yet unaffected. Such censoring can substantially bias estimates of correlation between relatives. To overcome this problem we develop a model for the association between liability to a disorder, and its age at onset. The model is designed for data from pairs of relatives to enable estimation of the correlation between an individuals’ liability to disorder and their age at onset. Usually, such information is not available at the individual level, because age at onset is uniquely available when onset has occurred. Lacking variation in disorder status, data from non-related persons cannot estimate the covariance between liability and age at onset. Data from relatives can resolve this issue when there is a correlation in liability between the relatives, because different age at onset distributions would be expected in concordant vs. discordant pairs of relatives. Greater severity and worse outcomes are often observed among those with earlier onset, so a correlation between disorder liability and age at onset seems likely in many cases. In this article we present the basic theory of the model, implemented as a mixture distribution, and an application to cannabis use in a Virginia Twin Study of Adolescent Behavioral Development. A negative association of (-.212) between age at onset an liability was found, with confidence intervals of −.263 to −.152, which do not cross zero. The method contrasts with Cox Proportional Hazards, in which disorder liability and onset timing are treated as a single dimension.

Estimates of per-individual genetic liability from large-scale family data are routinely used in biomedical research to describe genetic etiology of traits and disorders, boost the power of gene-mapping studies, and improve risk predictions. Here we present LT-FGRS, an R package for handling population-scale pedigrees and implementing multiple state-of-the-field methods for estimating genetic liability from such data. Benchmarking in population-scale Nordic registry data demonstrates that LT-FGRS reproduces estimates from existing implementations at manageable computational cost. LT-FGRS unifies previous parallel implementations into a single framework, lowering barriers for methodological comparison and applied use. Availability and Implementation: LT-FGRS is available as an R-package on CRAN. ( https://CRAN.R-project.org/package=LTFGRS ) Contact: emp@au.dk ; morten.dybdahl.krebs@regionh.dk Supplementary information: https://emilmip.github.io/LTFGRS

A bstract: Selective participation in biobanks often compromises inference to the general population, particularly when selection occurs across multiple stages, whether at recruitment or during subsequent participation. Inverse probability (IP) weighting can reduce systematic differences using suitable external benchmarks, but most applications assume a single selection process. Here, we present a multi-stage IP-weighting framework and apply it to the Danish Blood Donor Study (DBDS), a nationwide biobank embedded in Denmark’s blood-donation infrastructure. Using national registers, we estimated year-specific probabilities of (i) donation activity and (ii) DBDS enrolment conditional on donation activity, yielding two-stage inclusion weights for 169,893 participants. These weights reduced inclusion-associated imbalance across the 52 auxiliary variables in the probability models by 97.6% (median) and, despite strong health selection under donation-based recruitment, reduced relative-prevalence discrepancies across held-out prescription phenotypes by 69.7% (median). The effective sample size after weighting was 30,627 (18.0% of 169,893). Combining the inclusion weights with questionnaire-specific response weights across five DBDS questionnaires (>500 questions) produced the largest changes from unweighted to weighted responses for health behaviours and symptom severity, including tobacco and alcohol consumption, menstrual-pain severity, restless-legs severity, nocturia, sleep disturbance, and fatigue. These findings support multi-stage IP-weighting to improve population alignment in biobanks with staged selection.

Identifying significant associations between genetic loci and psychiatric disorders is dependent on very large sample sizes. Methods for diagnosing diseases on this scale, such as the use of self-assessment questionnaires and data from electronic health records, incorporate heritable variation unrelated to the disease of interest into the diagnosis. Consequently, genetic mapping will identify loci unrelated to the target disease while missing some that are related, and genetic correlations cannot be used to infer the genetic relationships between diseases and between cohorts. Furthermore, shared biases between different disorders appear as shared etiology. As sample sizes grow, such confounders propagate, and findings based on their presence are replicated and extended. Here, we draw attention to the problem, make suggestions for flagging affected cohorts, and discuss future data collection and machine learning approaches to mitigate the effects of heritable confounders in psychiatric disorders.

Importance: The incidence of ADHD and autism spectrum disorder (ASD) has increased markedly over recent decades, raising concerns about the emergence of new risk factors. Current literature typically attributes increased rates to changes in diagnostic practice, stigmatization, and awareness, but critically few studies have explored changes in underlying risk factors. Objective: To assess changes in the genetic risk profile of individuals diagnosed with ASD or ADHD according to year of incident diagnosis. Design: We used the iPSYCH2015 study, a population-based case-cohort with complete ascertainment of incident diagnoses for ASD and ADHD made from 1994 to 2016. Setting: Denmark. Participants: ASD (N=17,071) and ADHD (N=20,111). Exposure: Year of incident diagnosis. Regression models tested changes in the mean genetic risk profile of individuals diagnosed in each consecutive year (1994-2016), adjusting for age, sex, and ancestry. Main Outcomes: We used polygenic scores for psychiatric (ADHD, ASD, depression, bipolar, schizophrenia) and cognitive-behavioral (addiction, educational attainment, IQ, neuroticism, risk-taking) outcomes to capture the genetic risk profiles of diagnosed individuals. Results: A more recent ADHD diagnosis was associated (p<0.001) with less genetic risk for ADHD (β=-0.06 SD per 10 years) and other disorders (ASD, bipolar, schizophrenia). Similarly, a more recent ASD diagnosis was associated with less genetic risk for ASD (β=-0.07) and other disorders/traits (bipolar, schizophrenia, educational attainment). Conclusions and Relevance: Our novel approach suggests that over recent decades diagnostic practice around ADHD and ASD has evolved to capture a different profile of genetic risk. These findings support broadening diagnostic criteria as the explanation for the rise in incidence, with implications for understanding prevalence trends in relation to changes in risk factors and clinical practice. KEY POINTS: Question: Has the genetic risk profile of individuals diagnosed with autism spectrum disorder (ASD) and ADHD changed as diagnostic rates have risen? Findings: In the iPSYCH case-cohort study, we observed the genetic contributions to ASD and ADHD diagnoses have weakened over the past two decades. Meaning: Our results suggest that recent increase in ASD and ADHD diagnoses coincide with a broadening of diagnostic criteria.

Trait similarities between spouses are a key factor that shapes the landscape of complex human traits. The driving force behind the spousal correlations can increase the overall prevalence of disorders, influence occurrences of comorbidities and bias estimations of genetic architectures. However, there is a lack of large-scale studies examining cultural differences and generational trends in spousal correlations for psychiatric disorders. Focusing on three national registries, we performed a large-scale analysis on spousal correlations across nine psychiatric disorders. We obtained the trait correlations from five million spousal pairs in Taiwan and then compared them with estimates from the Danish national registry (571,534 pairs) and with published results from the Swedish national registry (707,263 pairs). Generational changes in Taiwan for people born after the 1930s were investigated as well. We found that a majority of psychiatric disorders have consistent spousal correlations across nations and over generations, indicating their importance in the population dynamics of psychiatric disorders.