Discovery of drug-omics associations in type 2 diabetes with generative deep-learning models.

The application of multiple omics technologies in biomedical cohorts has the potential to reveal patient-level disease characteristics and individualized response to treatment. However, the scale and heterogeneous nature of multi-modal data makes integration and inference a non-trivial task. We developed a deep-learning-based framework, multi-omics variational autoencoders (MOVE), to integrate such data and applied it to a cohort of 789 people with newly diagnosed type 2 diabetes with deep multi-omics phenotyping from the DIRECT consortium. Using in silico perturbations, we identified drug-omics associations across the multi-modal datasets for the 20 most prevalent drugs given to people with type 2 diabetes with substantially higher sensitivity than univariate statistical tests. From these, we among others, identified novel associations between metformin and the gut microbiota as well as opposite molecular responses for the two statins, simvastatin and atorvastatin. We used the associations to quantify drug-drug similarities, assess the degree of polypharmacy and conclude that drug effects are distributed across the multi-omics modalities.

An immunogenetic investigation of 30 autoimmune and autoinflammatory diseases and their links to psychiatric disorders in a nationwide sample.

Autoimmune and autoinflammatory diseases (AIIDs) involve a deficit in an individual's immune system function, whereby the immune reaction is directed against self-antigens. Many AIIDs have a strong genetic component, but they can also be triggered by environmental factors. AIIDs often have a highly negative impact on the individual's physical and mental wellbeing. Understanding the genetic underpinning of AIIDs is thus crucial both for diagnosis and for identifying individuals at high risk of an AIID and mental illness as a result thereof. The aim of the present study was to perform systematic statistical and genetic analyses to assess the role of human leukocyte antigen (HLA) alleles in 30 AIIDs and to study the links between AIIDs and psychiatric disorders. We leveraged the Danish iPSYCH Consortium sample comprising 65 534 individuals diagnosed with psychiatric disorders or selected as part of a random population sample, for whom we also had genetic data and diagnoses of AIIDs. We employed regression analysis to examine comorbidities between AIIDs and psychiatric disorders and associations between AIIDs and HLA alleles across seven HLA genes. Our comorbidity analyses showed that overall AIID and five specific AIIDs were associated with having a psychiatric diagnosis. Our genetic analyses found 81 significant associations between HLA alleles and AIIDs. Lastly, we show connections across AIIDs, psychiatric disorders and infection susceptibility through network analysis of significant HLA associations in these disease classes. Combined, our results include both novel associations as well as replications of previously reported associations in a large sample, and highlight the genetic and epidemiological links between AIIDs and psychiatric disorders.

Deep Learning for Cross-Diagnostic Prediction of Mental Disorder Diagnosis and Prognosis Using Danish Nationwide Register and Genetic Data.

Importance: Diagnoses and treatment of mental disorders are hampered by the current lack of objective markers needed to provide a more precise diagnosis and treatment strategy. Objective: To develop deep learning models to predict mental disorder diagnosis and severity spanning multiple diagnoses using nationwide register data, family and patient-specific diagnostic history, birth-related measurement, and genetics. Design, Setting, and Participants: This study was conducted from May 1, 1981, to December 31, 2016. For the analysis, which used a Danish population-based case-cohort sample of individuals born between 1981 and 2005, genotype data and matched longitudinal health register data were taken from the longitudinal Danish population-based Integrative Psychiatric Research Consortium 2012 case-cohort study. Included were individuals with mental disorders (attention-deficit/hyperactivity disorder [ADHD]), autism spectrum disorder (ASD), major depressive disorder (MDD), bipolar disorder (BD), schizophrenia spectrum disorders (SCZ), and population controls. Data were analyzed from February 1, 2021, to January 24, 2022. Exposure: At least 1 hospital contact with diagnosis of ADHD, ASD, MDD, BD, or SCZ. Main Outcomes and Measures: The predictability of (1) mental disorder diagnosis and (2) severity trajectories (measured by future outpatient hospital contacts, admissions, and suicide attempts) were investigated using both a cross-diagnostic and single-disorder setup. Predictive power was measured by AUC, accuracy, and Matthews correlation coefficient (MCC), including an estimate of feature importance. Results: A total of 63 535 individuals (mean [SD] age, 23 [7] years; 34 944 male [55%]; 28 591 female [45%]) were included in the model. Based on data prior to diagnosis, the specific diagnosis was predicted in a multidiagnostic prediction model including the background population with an overall area under the curve (AUC) of 0.81 and MCC of 0.28, whereas the single-disorder models gave AUCs/MCCs of 0.84/0.54 for SCZ, 0.79/0.41 for BD, 0.77/0.39 for ASD, 0.74/0.38, for ADHD, and 0.74/0.38 for MDD. The most important data sets for multidiagnostic prediction were previous mental disorders and age (11%-23% reduction in prediction accuracy when removed) followed by family diagnoses, birth-related measurements, and genetic data (3%-5% reduction in prediction accuracy when removed). Furthermore, when predicting subsequent disease trajectories of the disorder, the most severe cases were the most easily predictable, with an AUC of 0.72. Conclusions and Relevance: Results of this diagnostic study suggest the possibility of combining genetics and registry data to predict both mental disorder diagnosis and disorder progression in a clinically relevant, cross-diagnostic setting prior to clinical assessment.

Deep learning-based integration of genetics with registry data for stratification of schizophrenia and depression.

Currently, psychiatric diagnoses are, in contrast to most other medical fields, based on subjective symptoms and observable signs and call for new and improved diagnostics to provide the most optimal care. On the basis of a deep learning approach, we performed unsupervised patient stratification of 19,636 patients with depression [major depressive disorder (MDD)] and/or schizophrenia (SCZ) and 22,467 population controls from the iPSYCH2012 case cohort. We integrated data of disorder severity, history of mental disorders and disease comorbidities, genetics, and medical birth data. From this, we stratified the individuals in six and seven unique clusters for MDD and SCZ, respectively. When censoring data until diagnosis, we could predict MDD clusters with areas under the curve (AUCs) of 0.54 to 0.80 and SCZ clusters with AUCs of 0.71 to 0.86. Overall cases and controls could be predicted with an AUC of 0.81, illustrating the utility of data-driven subgrouping in psychiatry.

A large-scale investigation into the role of classical HLA loci in multiple types of severe infections, with a focus on overlaps with autoimmune and mental disorders.

BACKGROUND: Infections are a major disease burden worldwide. While they are caused by external pathogens, host genetics also plays a part in susceptibility to infections. Past studies have reported diverse associations between human leukocyte antigen (HLA) alleles and infections, but many were limited by small sample sizes and/or focused on only one infection. METHODS: We performed an immunogenetic association study examining 13 categories of severe infection (bacterial, viral, central nervous system, gastrointestinal, genital, hepatitis, otitis, pregnancy-related, respiratory, sepsis, skin infection, urological and other infections), as well as a phenotype for having any infection, and seven classical HLA loci (HLA-A, B, C, DPB1, DQA1, DQB1 and DRB1). Additionally, we examined associations between infections and specific alleles highlighted in our previous studies of psychiatric disorders and autoimmune disease, as these conditions are known to be linked to infections. RESULTS: Associations between HLA loci and infections were generally not strong. Highlighted associations included associations between DQB1*0302 and DQB1*0604 and viral infections (P = 0.002835 and P = 0.014332, respectively), DQB1*0503 and sepsis (P = 0.006053), and DQA1*0301 with "other" infections (a category which includes infections not included in our main categories e.g. protozoan infections) (P = 0.000369). Some HLA alleles implicated in autoimmune diseases showed association with susceptibility to infections, but the latter associations were generally weaker, or with opposite trends (in the case of HLA-C alleles, but not with alleles of HLA class II genes). HLA alleles associated with psychiatric disorders did not show association with susceptibility to infections. CONCLUSIONS: Our results suggest that classical HLA alleles do not play a large role in the etiology of severe infections. The discordant association trends with autoimmune disease for some alleles could contribute to mechanistic theories of disease etiology.

Improved metagenome binning and assembly using deep variational autoencoders.

Despite recent advances in metagenomic binning, reconstruction of microbial species from metagenomics data remains challenging. Here we develop variational autoencoders for metagenomic binning (VAMB), a program that uses deep variational autoencoders to encode sequence coabundance and k-mer distribution information before clustering. We show that a variational autoencoder is able to integrate these two distinct data types without any previous knowledge of the datasets. VAMB outperforms existing state-of-the-art binners, reconstructing 29-98% and 45% more near-complete (NC) genomes on simulated and real data, respectively. Furthermore, VAMB is able to separate closely related strains up to 99.5% average nucleotide identity (ANI), and reconstructed 255 and 91 NC Bacteroides vulgatus and Bacteroides dorei sample-specific genomes as two distinct clusters from a dataset of 1,000 human gut microbiome samples. We use 2,606 NC bins from this dataset to show that species of the human gut microbiome have different geographical distribution patterns. VAMB can be run on standard hardware and is freely available at https://github.com/RasmussenLab/vamb .

Immunity and mental illness: findings from a Danish population-based immunogenetic study of seven psychiatric and neurodevelopmental disorders.

Human leukocyte antigen (HLA) genes encode proteins with important roles in the regulation of the immune system. Many studies have also implicated HLA genes in psychiatric and neurodevelopmental disorders. However, these studies usually focus on one disorder and/or on one HLA candidate gene, often with small samples. Here, we access a large dataset of 65,534 genotyped individuals consisting of controls (N = 19,645) and cases having one or more of autism spectrum disorder (N = 12,331), attention deficit hyperactivity disorder (N = 14,397), schizophrenia (N = 2401), bipolar disorder (N = 1391), depression (N = 18,511), anorexia (N = 2551) or intellectual disability (N = 3175). We imputed participants' HLA alleles to investigate the involvement of HLA genes in these disorders using regression models. We found a pronounced protective effect of DPB1*1501 on susceptibility to autism (p = 0.0094, OR = 0.72) and intellectual disability (p = 0.00099, OR = 0.41), with an increased protective effect on a comorbid diagnosis of both disorders (p = 0.003, OR = 0.29). We also identified a risk allele for intellectual disability, B*5701 (p = 0.00016, OR = 1.33). Associations with both alleles survived FDR correction and a permutation procedure. We did not find significant evidence for replication of previously-reported associations for autism or schizophrenia. Our results support an implication of HLA genes in autism and intellectual disability, which requires replication by other studies. Our study also highlights the importance of large sample sizes in HLA association studies.

SCCmecFinder, a Web-Based Tool for Typing of Staphylococcal Cassette Chromosome mec in Staphylococcus aureus Using Whole-Genome Sequence Data.

Typing of methicillin-resistant Staphylococcus aureus (MRSA) is important in infection control and surveillance. The current nomenclature of MRSA includes the genetic background of the S. aureus strain determined by multilocus sequence typing (MLST) or equivalent methods like spa typing and typing of the mobile genetic element staphylococcal cassette chromosome mec (SCCmec), which carries the mecA or mecC gene. Whereas MLST and spa typing are relatively simple, typing of SCCmec is less trivial because of its heterogeneity. Whole-genome sequencing (WGS) provides the essential data for typing of the genetic background and SCCmec, but so far, no bioinformatic tools for SCCmec typing have been available. Here, we report the development and evaluation of SCCmecFinder for characterization of the SCCmec element from S. aureus WGS data. SCCmecFinder is able to identify all SCCmec element types, designated I to XIII, with subtyping of SCCmec types IV (2B) and V (5C2). SCCmec elements are characterized by two different gene prediction approaches to achieve correct annotation, a Basic Local Alignment Search Tool (BLAST)-based approach and a k-mer-based approach. Evaluation of SCCmecFinder by using a diverse collection of clinical isolates (n = 93) showed a high typeability level of 96.7%, which increased to 98.9% upon modification of the default settings. In conclusion, SCCmecFinder can be an alternative to more laborious SCCmec typing methods and is freely available at https://cge.cbs.dtu.dk/services/SCCmecFinder. IMPORTANCE SCCmec in MRSA is acknowledged to be of importance not only because it contains the mecA or mecC gene but also for staphylococcal adaptation to different environments, e.g., in hospitals, the community, and livestock. Typing of SCCmec by PCR techniques has, because of its heterogeneity, been challenging, and whole-genome sequencing has only partially solved this since no good bioinformatic tools have been available. In this article, we describe the development of a new bioinformatic tool, SCCmecFinder, that includes most of the needs for infection control professionals and researchers regarding the interpretation of SCCmec elements. The software detects all of the SCCmec elements accepted by the International Working Group on the Classification of Staphylococcal Cassette Chromosome Elements, and users will be prompted if diverging and potential new elements are uploaded. Furthermore, SCCmecFinder will be curated and updated as new elements are found and it is easy to use and freely accessible.

Direct whole-genome sequencing of Plasmodium falciparum specimens from dried erythrocyte spots.

BACKGROUND: Plasmodium falciparum malaria remains a major health burden and genomic research represents one of the necessary approaches for continued progress towards malaria control and elimination. Sample acquisition for this purpose is troublesome, with the majority of malaria-infected individuals living in rural areas, away from main infrastructure and the electrical grid. The aim of this study was to describe a low-tech procedure to sample P. falciparum specimens for direct whole genome sequencing (WGS), without use of electricity and cold-chain. METHODS: Venous blood samples were collected from malaria patients in Bandim, Guinea-Bissau and leukocyte-depleted using Plasmodipur filters, the enriched parasite sample was spotted on Whatman paper and dried. The samples were stored at ambient temperatures and subsequently used for DNA-extraction. Ratios of parasite:human content of the extracted DNA was assessed by qPCR, and five samples with varying parasitaemia, were sequenced. Sequencing data were used to analyse the sample content, as well as sample coverage and depth as compared to the 3d7 reference genome. RESULTS: qPCR revealed that 73% of the 199 samples were applicable for WGS, as defined by a minimum ratio of parasite:human DNA of 2:1. WGS revealed an even distribution of sequence data across the 3d7 reference genome, regardless of parasitaemia. The acquired read depths varied from 16 to 99×, and coverage varied from 87.5 to 98.9% of the 3d7 reference genome. SNP-analysis of six genes, for which amplicon sequencing has been performed previously, confirmed the reliability of the WGS-data. CONCLUSION: This study describes a simple filter paper based protocol for sampling P. falciparum from malaria patients for subsequent direct WGS, enabling acquisition of samples in remote settings with no access to electricity.