Genome-wide Association Studies in Ancestrally Diverse Populations: Opportunities, Methods, Pitfalls, and Recommendations.

Genome-wide association studies (GWASs) have focused primarily on populations of European descent, but it is essential that diverse populations become better represented. Increasing diversity among study participants will advance our understanding of genetic architecture in all populations and ensure that genetic research is broadly applicable. To facilitate and promote research in multi-ancestry and admixed cohorts, we outline key methodological considerations and highlight opportunities, challenges, solutions, and areas in need of development. Despite the perception that analyzing genetic data from diverse populations is difficult, it is scientifically and ethically imperative, and there is an expanding analytical toolbox to do it well.

BioM2: biologically informed multi-stage machine learning for phenotype prediction using omics data.

Navigating the complex landscape of high-dimensional omics data with machine learning models presents a significant challenge. The integration of biological domain knowledge into these models has shown promise in creating more meaningful stratifications of predictor variables, leading to algorithms that are both more accurate and generalizable. However, the wider availability of machine learning tools capable of incorporating such biological knowledge remains limited. Addressing this gap, we introduce BioM2, a novel R package designed for biologically informed multistage machine learning. BioM2 uniquely leverages biological information to effectively stratify and aggregate high-dimensional biological data in the context of machine learning. Demonstrating its utility with genome-wide DNA methylation and transcriptome-wide gene expression data, BioM2 has shown to enhance predictive performance, surpassing traditional machine learning models that operate without the integration of biological knowledge. A key feature of BioM2 is its ability to rank predictor variables within biological categories, specifically Gene Ontology pathways. This functionality not only aids in the interpretability of the results but also enables a subsequent modular network analysis of these variables, shedding light on the intricate systems-level biology underpinning the predictive outcome. We have proposed a biologically informed multistage machine learning framework termed BioM2 for phenotype prediction based on omics data. BioM2 has been incorporated into the BioM2 CRAN package (https://cran.r-project.org/web/packages/BioM2/index.html).

A primer on the use of machine learning to distil knowledge from data in biological psychiatry.

Applications of machine learning in the biomedical sciences are growing rapidly. This growth has been spurred by diverse cross-institutional and interdisciplinary collaborations, public availability of large datasets, an increase in the accessibility of analytic routines, and the availability of powerful computing resources. With this increased access and exposure to machine learning comes a responsibility for education and a deeper understanding of its bases and bounds, borne equally by data scientists seeking to ply their analytic wares in medical research and by biomedical scientists seeking to harness such methods to glean knowledge from data. This article provides an accessible and critical review of machine learning for a biomedically informed audience, as well as its applications in psychiatry. The review covers definitions and expositions of commonly used machine learning methods, and historical trends of their use in psychiatry. We also provide a set of standards, namely Guidelines for REporting Machine Learning Investigations in Neuropsychiatry (GREMLIN), for designing and reporting studies that use machine learning as a primary data-analysis approach. Lastly, we propose the establishment of the Machine Learning in Psychiatry (MLPsych) Consortium, enumerate its objectives, and identify areas of opportunity for future applications of machine learning in biological psychiatry. This review serves as a cautiously optimistic primer on machine learning for those on the precipice as they prepare to dive into the field, either as methodological practitioners or well-informed consumers.

Knowledge graphs in psychiatric research: Potential applications and future perspectives.

BACKGROUND: Knowledge graphs (KGs) remain an underutilized tool in the field of psychiatric research. In the broader biomedical field KGs are already a significant tool mainly used as knowledge database or for novel relation detection between biomedical entities. This review aims to outline how KGs would further research in the field of psychiatry in the age of Artificial Intelligence (AI) and Large Language Models (LLMs). METHODS: We conducted a thorough literature review across a spectrum of scientific fields ranging from computer science and knowledge engineering to bioinformatics. The literature reviewed was taken from PubMed, Semantic Scholar and Google Scholar searches including terms such as "Psychiatric Knowledge Graphs", "Biomedical Knowledge Graphs", "Knowledge Graph Machine Learning Applications", "Knowledge Graph Applications for Biomedical Sciences". The resulting publications were then assessed and accumulated in this review regarding their possible relevance to future psychiatric applications. RESULTS: A multitude of papers and applications of KGs in associated research fields that are yet to be utilized in psychiatric research was found and outlined in this review. We create a thorough recommendation for other computational researchers regarding use-cases of these KG applications in psychiatry. CONCLUSION: This review illustrates use-cases of KG-based research applications in biomedicine and beyond that may aid in elucidating the complex biology of psychiatric illness and open new routes for developing innovative interventions. We conclude that there is a wealth of opportunities for KG utilization in psychiatric research across a variety of application areas including biomarker discovery, patient stratification and personalized medicine approaches.

Deciphering the interplay between psychopathological symptoms, sensorimotor, cognitive and global functioning: a transdiagnostic network analysis.

BACKGROUND: Understanding the relationship between psychopathology and major domains of human neurobehavioral functioning may identify new transdiagnostic treatment targets. However, studies examining the interrelationship between psychopathological symptoms, sensorimotor, cognitive, and global functioning in a transdiagnostic sample are lacking. We hypothesized a close relationship between sensorimotor and cognitive functioning in a transdiagnostic patient sample. METHODS: We applied network analysis and community detection methods to examine the interplay and centrality [expected influence (EI) and strength] between psychopathological symptoms, sensorimotor, cognitive, and global functioning in a transdiagnostic sample consisting of 174 schizophrenia spectrum (SSD) and 38 mood disorder (MOD) patients. All patients (n = 212) were examined with the Positive and Negative Syndrome Scale (PANSS), the Heidelberg Neurological Soft Signs Scale (NSS), the Global Assessment of Functioning (GAF), and the Brief Cognitive Assessment Tool for Schizophrenia consisted of trail making test B (TMT-B), category fluency (CF) and digit symbol substitution test (DSST). RESULTS: NSS showed closer connections with TMT-B, CF, and DSST than with GAF and PANSS. DSST, PANSS general, and NSS motor coordination scores showed the highest EI. Sensory integration, DSST, and CF showed the highest strength. CONCLUSIONS: The close connection between sensorimotor and cognitive impairment as well as the high centrality of sensorimotor symptoms suggests that both domains share aspects of SSD and MOD pathophysiology. But, because the majority of the study population was diagnosed with SSD, the question as to whether sensorimotor symptoms are really a transdiagnostic therapeutic target needs to be examined in future studies including more balanced diagnostic groups.

The ReCoDe addiction research consortium: Losing and regaining control over drug intake-Findings and future perspectives.

Substance use disorders (SUDs) are seen as a continuum ranging from goal-directed and hedonic drug use to loss of control over drug intake with aversive consequences for mental and physical health and social functioning. The main goals of our interdisciplinary German collaborative research centre on Losing and Regaining Control over Drug Intake (ReCoDe) are (i) to study triggers (drug cues, stressors, drug priming) and modifying factors (age, gender, physical activity, cognitive functions, childhood adversity, social factors, such as loneliness and social contact/interaction) that longitudinally modulate the trajectories of losing and regaining control over drug consumption under real-life conditions. (ii) To study underlying behavioural, cognitive and neurobiological mechanisms of disease trajectories and drug-related behaviours and (iii) to provide non-invasive mechanism-based interventions. These goals are achieved by: (A) using innovative mHealth (mobile health) tools to longitudinally monitor the effects of triggers and modifying factors on drug consumption patterns in real life in a cohort of 900 patients with alcohol use disorder. This approach will be complemented by animal models of addiction with 24/7 automated behavioural monitoring across an entire disease trajectory; i.e. from a naïve state to a drug-taking state to an addiction or resilience-like state. (B) The identification and, if applicable, computational modelling of key molecular, neurobiological and psychological mechanisms (e.g., reduced cognitive flexibility) mediating the effects of such triggers and modifying factors on disease trajectories. (C) Developing and testing non-invasive interventions (e.g., Just-In-Time-Adaptive-Interventions (JITAIs), various non-invasive brain stimulations (NIBS), individualized physical activity) that specifically target the underlying mechanisms for regaining control over drug intake. Here, we will report on the most important results of the first funding period and outline our future research strategy.

Network-based artificial intelligence approaches for advancing personalized psychiatry.

Psychiatric disorders have a complex biological underpinning likely involving an interplay of genetic and environmental risk contributions. Substantial efforts are being made to use artificial intelligence approaches to integrate features within and across data types to increase our etiological understanding and advance personalized psychiatry. Network science offers a conceptual framework for exploring the often complex relationships across different levels of biological organization, from cellular mechanistic to brain-functional and phenotypic networks. Utilizing such network information effectively as part of artificial intelligence approaches is a promising route toward a more in-depth understanding of illness biology, the deciphering of patient heterogeneity, and the identification of signatures that may be sufficiently predictive to be clinically useful. Here, we present examples of how network information has been used as part of artificial intelligence within psychiatry and beyond and outline future perspectives on how personalized psychiatry approaches may profit from a closer integration of psychiatric research, artificial intelligence development, and network science.

dsMTL: a computational framework for privacy-preserving, distributed multi-task machine learning.

MOTIVATION: In multi-cohort machine learning studies, it is critical to differentiate between effects that are reproducible across cohorts and those that are cohort-specific. Multi-task learning (MTL) is a machine learning approach that facilitates this differentiation through the simultaneous learning of prediction tasks across cohorts. Since multi-cohort data can often not be combined into a single storage solution, there would be the substantial utility of an MTL application for geographically distributed data sources. RESULTS: Here, we describe the development of 'dsMTL', a computational framework for privacy-preserving, distributed multi-task machine learning that includes three supervised and one unsupervised algorithms. First, we derive the theoretical properties of these methods and the relevant machine learning workflows to ensure the validity of the software implementation. Second, we implement dsMTL as a library for the R programming language, building on the DataSHIELD platform that supports the federated analysis of sensitive individual-level data. Third, we demonstrate the applicability of dsMTL for comorbidity modeling in distributed data. We show that comorbidity modeling using dsMTL outperformed conventional, federated machine learning, as well as the aggregation of multiple models built on the distributed datasets individually. The application of dsMTL was computationally efficient and highly scalable when applied to moderate-size (n < 500), real expression data given the actual network latency. AVAILABILITY AND IMPLEMENTATION: dsMTL is freely available at https://github.com/transbioZI/dsMTLBase (server-side package) and https://github.com/transbioZI/dsMTLClient (client-side package). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Behavioural and functional evidence revealing the role of RBFOX1 variation in multiple psychiatric disorders and traits.

Common variation in the gene encoding the neuron-specific RNA splicing factor RNA Binding Fox-1 Homolog 1 (RBFOX1) has been identified as a risk factor for several psychiatric conditions, and rare genetic variants have been found causal for autism spectrum disorder (ASD). Here, we explored the genetic landscape of RBFOX1 more deeply, integrating evidence from existing and new human studies as well as studies in Rbfox1 knockout mice. Mining existing data from large-scale studies of human common genetic variants, we confirmed gene-based and genome-wide association of RBFOX1 with risk tolerance, major depressive disorder and schizophrenia. Data on six mental disorders revealed copy number losses and gains to be more frequent in ASD cases than in controls. Consistently, RBFOX1 expression appeared decreased in post-mortem frontal and temporal cortices of individuals with ASD and prefrontal cortex of individuals with schizophrenia. Brain-functional MRI studies demonstrated that carriers of a common RBFOX1 variant, rs6500744, displayed increased neural reactivity to emotional stimuli, reduced prefrontal processing during cognitive control, and enhanced fear expression after fear conditioning, going along with increased avoidance behaviour. Investigating Rbfox1 neuron-specific knockout mice allowed us to further specify the role of this gene in behaviour. The model was characterised by pronounced hyperactivity, stereotyped behaviour, impairments in fear acquisition and extinction, reduced social interest, and lack of aggression; it provides excellent construct and face validity as an animal model of ASD. In conclusion, convergent translational evidence shows that common variants in RBFOX1 are associated with a broad spectrum of psychiatric traits and disorders, while rare genetic variation seems to expose to early-onset neurodevelopmental psychiatric disorders with and without developmental delay like ASD, in particular. Studying the pleiotropic nature of RBFOX1 can profoundly enhance our understanding of mental disorder vulnerability.

Genetic control of variability in subcortical and intracranial volumes.

Sensitivity to external demands is essential for adaptation to dynamic environments, but comes at the cost of increased risk of adverse outcomes when facing poor environmental conditions. Here, we apply a novel methodology to perform genome-wide association analysis of mean and variance in ten key brain features (accumbens, amygdala, caudate, hippocampus, pallidum, putamen, thalamus, intracranial volume, cortical surface area, and cortical thickness), integrating genetic and neuroanatomical data from a large lifespan sample (n = 25,575 individuals; 8-89 years, mean age 51.9 years). We identify genetic loci associated with phenotypic variability in thalamus volume and cortical thickness. The variance-controlling loci involved genes with a documented role in brain and mental health and were not associated with the mean anatomical volumes. This proof-of-principle of the hypothesis of a genetic regulation of brain volume variability contributes to establishing the genetic basis of phenotypic variance (i.e., heritability), allows identifying different degrees of brain robustness across individuals, and opens new research avenues in the search for mechanisms controlling brain and mental health.

Association of polygenic risk for schizophrenia with fast sleep spindle density depends on pro-cognitive variants.

Cognitive impairment is a common feature in schizophrenia and the strongest prognostic factor for long-term outcome. Identifying a trait associated with the genetic background for cognitive outcome in schizophrenia may aid in a deeper understanding of clinical disease subtypes. Fast sleep spindles may represent such a biomarker as they are strongly genetically determined, associated with cognitive functioning and impaired in schizophrenia and unaffected relatives. We measured fast sleep spindle density in 150 healthy adults and investigated its association with a genome-wide polygenic score for schizophrenia (SCZ-PGS). The association between SCZ-PGS and fast spindle density was further characterized by stratifying it to the genetic background of intelligence. SCZ-PGS was positively associated with fast spindle density. This association mainly depended on pro-cognitive genetic variants. Our results strengthen the evidence for a genetic background of spindle abnormalities in schizophrenia. Spindle density might represent an easily accessible marker for a favourable cognitive outcome which should be further investigated in clinical samples.

Brain structural correlates of upward social mobility in ethnic minority individuals.

PURPOSE: Perigenual anterior cingulate cortex (pACC) is a neural convergence site for social stress-related risk factors for mental health, including ethnic minority status. Current social status, a strong predictor of mental and somatic health, has been related to gray matter volume in this region, but the effects of social mobility over the lifespan are unknown and may differ in minorities. Recent studies suggest a diminished health return of upward social mobility for ethnic minority individuals, potentially due to sustained stress-associated experiences and subsequent activation of the neural stress response system. METHODS: To address this issue, we studied an ethnic minority sample with strong upward social mobility. In a cross-sectional design, we examined 64 young adult native German and 76 ethnic minority individuals with comparable sociodemographic attributes using whole-brain structural magnetic resonance imaging. RESULTS: Results showed a significant group-dependent interaction between perceived upward social mobility and pACC gray matter volume, with a significant negative association in the ethnic minority individuals. Post-hoc analysis showed a significant mediation of the relationship between perceived upward social mobility and pACC volume by perceived chronic stress, a variable that was significantly correlated with perceived discrimination in our ethnic minority group. CONCLUSION: Our findings extend prior work by pointing to a biological signature of the "allostatic costs" of socioeconomic attainment in socially disadvantaged upwardly mobile individuals in a key neural node implicated in the regulation of stress and negative affect.

Differential resting-state patterns across networks are spatially associated with Comt and Trmt2a gene expression patterns in a mouse model of 22q11.2 deletion.

Copy number variations (CNV) involving multiple genes are ideal models to study polygenic neuropsychiatric disorders. Since 22q11.2 deletion is regarded as the most important single genetic risk factor for developing schizophrenia, characterizing the effects of this CNV on neural networks offers a unique avenue towards delineating polygenic interactions conferring risk for the disorder. We used a Df(h22q11)/+ mouse model of human 22q11.2 deletion to dissect gene expression patterns that would spatially overlap with differential resting-state functional connectivity (FC) patterns in this model (N = 12 Df(h22q11)/+ mice, N = 10 littermate controls). To confirm the translational relevance of our findings, we analyzed tissue samples from schizophrenia patients and healthy controls using machine learning to explore whether identified genes were co-expressed in humans. Additionally, we employed the STRING protein-protein interaction database to identify potential interactions between genes spatially associated with hypo- or hyper-FC. We found significant associations between differential resting-state connectivity and spatial gene expression patterns for both hypo- and hyper-FC. Two genes, Comt and Trmt2a, were consistently over-expressed across all networks. An analysis of human datasets pointed to a disrupted co-expression of these two genes in the brain in schizophrenia patients, but not in healthy controls. Our findings suggest that COMT and TRMT2A form a core genetic component implicated in differential resting-state connectivity patterns in the 22q11.2 deletion. A disruption of their co-expression in schizophrenia patients points out a prospective cause for the aberrance of brain networks communication in 22q11.2 deletion syndrome on a molecular level.

Generative network models of altered structural brain connectivity in schizophrenia.

Alterations in the structural connectome of schizophrenia patients have been widely characterized, but the mechanisms remain largely unknown. Generative network models have recently been introduced as a tool to test the biological underpinnings of altered brain network formation. We evaluated different generative network models in healthy controls (n=152), schizophrenia patients (n=66), and their unaffected first-degree relatives (n=32), and we identified spatial and topological factors contributing to network formation. We further investigated how these factors relate to cognition and to polygenic risk for schizophrenia. Our data show that among the four tested classes of generative network models, structural brain networks were optimally accounted for by a two-factor model combining spatial constraints and topological neighborhood structure. The same wiring model explained brain network formation across study groups. However, relatives and schizophrenia patients exhibited significantly lower spatial constraints and lower topological facilitation compared to healthy controls. Further exploratory analyses point to potential associations of the model parameter reflecting spatial constraints with the polygenic risk for schizophrenia and cognitive performance. Our results identify spatial constraints and local topological structure as two interrelated mechanisms contributing to regular brain network formation as well as altered connectomes in schizophrenia and healthy individuals at familial risk for schizophrenia. On an exploratory level, our data further point to the potential relevance of spatial constraints for the genetic risk for schizophrenia and general cognitive functioning, thereby encouraging future studies in following up on these observations to gain further insights into the biological basis and behavioral relevance of model parameters.

Multimodal imaging improves brain age prediction and reveals distinct abnormalities in patients with psychiatric and neurological disorders.

The deviation between chronological age and age predicted using brain MRI is a putative marker of overall brain health. Age prediction based on structural MRI data shows high accuracy in common brain disorders. However, brain aging is complex and heterogenous, both in terms of individual differences and the underlying biological processes. Here, we implemented a multimodal model to estimate brain age using different combinations of cortical area, thickness and sub-cortical volumes, cortical and subcortical T1/T2-weighted ratios, and cerebral blood flow (CBF) based on arterial spin labeling. For each of the 11 models we assessed the age prediction accuracy in healthy controls (HC, n = 750) and compared the obtained brain age gaps (BAGs) between age-matched subsets of HC and patients with Alzheimer's disease (AD, n = 54), mild (MCI, n = 90) and subjective (SCI, n = 56) cognitive impairment, schizophrenia spectrum (SZ, n = 159) and bipolar disorder (BD, n = 135). We found highest age prediction accuracy in HC when integrating all modalities. Furthermore, two-group case-control classifications revealed highest accuracy for AD using global T1-weighted BAG, while MCI, SCI, BD and SZ showed strongest effects in CBF-based BAGs. Combining multiple MRI modalities improves brain age prediction and reveals distinct deviations in patients with psychiatric and neurological disorders. The multimodal BAG was most accurate in predicting age in HC, while group differences between patients and HC were often larger for BAGs based on single modalities. These findings indicate that multidimensional neuroimaging of patients may provide a brain-based mapping of overlapping and distinct pathophysiology in common disorders.

[Twelve years of research domain criteria in psychiatric research and practice: claim and reality]. / Zwölf Jahre Research Domain Criteria in der psychiatrischen Forschung und Praxis: Anspruch und Wirklichkeit.

The research domain criteria (RDoC) initiative of the National Institute of Mental Health (NIMH) was presented 12 years ago. The RDoC provides a matrix for the systematic, dimensional and domain-based study of mental disorders that is not based on established disease entities as defined by the Diagnostic and Statistical Manual of Mental Disorders (DSM) or the International Classification of Diseases (ICD). The primary aim of RDoC is to understand the nature of mental health and illness in terms of different extents of dysfunction in psychological/biological systems with interconnected diagnoses. This selective review article aims to provide a comprehensive overview of RDoC-based studies that have contributed to a better conceptual organization of mental disorders. Numerous promising and methodologically sophisticated studies on RDoC were identified. The number of scientific studies increased over time, indicating that dimensional research is increasingly being pursued in psychiatry. In summary, the RDoC initiative has a considerable potential to more precisely define the complexity of pathomechanisms underlying mental disorders; however, major challenges (e.g. small and heterogeneous study samples, unclear biomarker definitions and lack of replication studies) remain to be overcome in the future. Furthermore, it is plausible that a diagnostic system of the future will integrate categorical and dimensional approaches to arrive at a stratification that can underpin a precision medical approach in psychiatry.

RMTL: an R library for multi-task learning.

MOTIVATION: Multi-task learning (MTL) is a machine learning technique for simultaneous learning of multiple related classification or regression tasks. Despite its increasing popularity, MTL algorithms are currently not available in the widely used software environment R, creating a bottleneck for their application in biomedical research. RESULTS: We developed an efficient, easy-to-use R library for MTL (www.r-project.org) comprising 10 algorithms applicable for regression, classification, joint predictor selection, task clustering, low-rank learning and incorporation of biological networks. We demonstrate the utility of the algorithms using simulated data. AVAILABILITY AND IMPLEMENTATION: The RMTL package is an open source R package and is freely available at https://github.com/transbioZI/RMTL. RMTL will also be available on cran.r-project.org. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Gimpute: an efficient genetic data imputation pipeline.

MOTIVATION: Genotype imputation is essential for genome-wide association studies (GWAS) to retrieve information of untyped variants and facilitate comparability across studies. However, there is a lack of automated pipelines that perform all required processing steps prior to and following imputation. RESULTS: Based on widely used and freely available tools, we have developed Gimpute, an automated processing and imputation pipeline for genome-wide association data. Gimpute includes processing steps for genotype liftOver, quality control, population outlier detection, haplotype pre-phasing, imputation, post imputation, data management and the extension to other existing pipeline. AVAILABILITY AND IMPLEMENTATION: The Gimpute package is an open source R package and is freely available at https://github.com/transbioZI/Gimpute. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Computational Approaches for Identification of Pleiotropic Biomarker Profiles in Psychiatry.

The discovery of biomarkers is considered a critical step towards an improved clinical management of psychiatric disorders. Despite the availability of advanced computational approaches, the lack of strong individual predictors of clinically relevant outcomes, combined with the usually high dimensionality, significantly hamper the identification of such markers. Consistent with the often observed lack of diagnostic specificity of biological alterations, research suggests an underlying genetic pleiotropy between psychiatric illnesses and frequently comorbid conditions, such as type 2 diabetes or cardiovascular illnesses. As research is transitioning away from conventional diagnostic delineations towards a dimensional understanding of psychiatric illness, gaining insight into such pleiotropy and its downstream biological effects bears promise for identification of clinically useful biomarkers. In this review, we summarize the computational methods for identifying biological markers indexing pleiotropic effects and discuss recent research findings in this context.