MELLODDY: Cross-pharma Federated Learning at Unprecedented Scale Unlocks Benefits in QSAR without Compromising Proprietary Information

Federated multipartner machine learning has been touted as an appealing and efficient method to increase the effective training data volume and thereby the predictivity of models, particularly when the generation of training data is resource-intensive. In the landmark MELLODDY project, indeed, each of ten pharmaceutical companies realized aggregated improvements on its own classification or regression models through federated learning. To this end, they leveraged a novel implementation extending multitask learning across partners, on a platform audited for privacy and security. The experiments involved an unprecedented cross-pharma data set of 2.6+ billion confidential experimental activity data points, documenting 21+ million physical small molecules and 40+ thousand assays in on-target and secondary pharmacodynamics and pharmacokinetics. Appropriate complementary metrics were developed to evaluate the predictive performance in the federated setting. In addition to predictive performance increases in labeled space, the results point toward an extended applicability domain in federated learning. Increases in collective training data volume, including by means of auxiliary data resulting from single concentration high-throughput and imaging assays, continued to boost predictive performance, albeit with a saturating return. Markedly higher improvements were observed for the pharmacokinetics and safety panel assay-based task subsets.

Olfactory genes affect major depression in highly educated, emotionally stable, lean women: a bridge between animal models and precision medicine

Most current approaches to establish subgroups of depressed patients for precision medicine aim to rely on biomarkers that require highly specialized assessment. Our present aim was to stratify participants of the UK Biobank cohort based on three readily measurable common independent risk factors, and to investigate depression genomics in each group to discover common and separate biological etiology. Two-step cluster analysis was run separately in males (n = 149,879) and females (n = 174,572), with neuroticism (a tendency to experience negative emotions), body fat percentage, and years spent in education as input variables. Genome-wide association analyses were implemented within each of the resulting clusters, for the lifetime occurrence of either a depressive episode or recurrent depressive disorder as the outcome. Variant-based, gene-based, gene set-based, and tissue-specific gene expression test were applied. Phenotypically distinct clusters with high genetic intercorrelations in depression genomics were found. A two-cluster solution was the best model in each sex with some differences including the less important role of neuroticism in males. In females, in case of a protective pattern of low neuroticism, low body fat percentage, and high level of education, depression was associated with pathways related to olfactory function. While also in females but in a risk pattern of high neuroticism, high body fat percentage, and less years spent in education, depression showed association with complement system genes. Our results, on one hand, indicate that alteration of olfactory pathways, that can be paralleled to the well-known rodent depression models of olfactory bulbectomy, might be a novel target towards precision psychiatry in females with less other risk factors for depression. On the other hand, our results in multi-risk females may provide a special case of immunometabolic depression.

Partial Thermal Condensation Mediated Synthesis of High-Density Nickel Single Atom Sites on Carbon Nitride for Selective Photooxidation of Methane into Methanol

Direct selective transformation of greenhouse methane (CH4) to liquid oxygenates (methanol) can substitute energy-intensive two-step (reforming/Fischer-Tropsch) synthesis while creating environmental benefits. The development of inexpensive, selective, and robust catalysts that enable room temperature conversion will decide the future of this technology. Single-atom catalysts (SACs) with isolated active centers embedded in support have displayed significant promises in catalysis to drive challenging reactions. Herein, high-density Ni single atoms are developed and stabilized on carbon nitride (NiCN) via thermal condensation of preorganized Ni-coordinated melem units. The physicochemical characterization of NiCN with various analytical techniques including HAADF-STEM and X-ray absorption fine structure (XAFS) validate the successful formation of Ni single atoms coordinated to the heptazine-constituted CN network. The presence of uniform catalytic sites improved visible absorption and carrier separation in densely populated NiCN SAC resulting in 100% selective photoconversion of (CH4) to methanol using H2O2 as an oxidant. The superior catalytic activity can be attributed to the generation of high oxidation (NiIII═O) sites and selective C─H bond cleavage to generate •CH3 radicals on Ni centers, which can combine with •OH radicals to generate CH3OH.

A sound mind in a sound body: a novel concept unravelling heterogeneity of depression

Depression is a highly prevalent and debilitating condition, yet we still lack both in-depth knowledge concerning its etiopathology and sufficiently efficacious treatment options. With approximately one third of patients resistant to currently available antidepressants there is a pressing need for a better understanding of depression, identifying subgroups within the highly heterogeneous illness category and to understand the divergent underlying biology of such subtypes, to help develop and personalise treatments. The TRAJECTOME project aims to address such challenges by (1) identifying depression-related multimorbidity subgroups and shared molecular pathways based on temporal disease profiles from healthcare systems and biobank data using machine learning approaches, and by (2) characterising these subgroups from multiple aspects including genetic variants, metabolic processes, lifestyle and environmental factors. Following the identification of multimorbidity trajectories, a disease burden score related to depression and adjusted for multimorbidity was established summarising the current state of the patient to weigh the molecular mechanisms associated with depression. In addition, the role of genetic and environmental factors, and also their interactions were identified for all subgroups. The project also attempted to identify potential metabolomic markers for the early diagnostics of these multimorbidity conditions. Finally, we prioritized molecular drug candidates matching the multimorbidity pathways indicated for the individual subgroups which would potentially offer personalised treatment simultaneously for the observable multimorbid conditions yet minimising polypharmacy and related side effects. The present paper overviews the TRAJECTOME project including its aims, tasks, procedures and accomplishments. (Neuropsychopharmacol Hung 2023; 25(4): 183-193)

Synergistic antibacterial action of the iron complex and ampicillin against Staphylococcus aureus

OBJECTIVES

Resistance to antibiotics among bacteria of clinical importance, including Staphylococcus aureus, is a serious problem worldwide and the search for alternatives is needed. Some metal complexes have antibacterial properties and when combined with antibiotics, they may increase bacterial sensitivity to antimicrobials. In this study, we synthesized the iron complex and tested it in combination with ampicillin (Fe16 + AMP) against S. aureus.

METHODS

An iron complex (Fe16) was synthesized and characterized using spectroscopy methods. Confirmation of the synergistic effect between the iron complex (Fe16) and ampicillin (AMP) was performed using ζ-potential, infrared spectra and FICI index calculated from the minimum inhibitory concentration (MIC) from the checkerboard assay. Cytotoxic properties of combination Fe16 + AMP was evaluated on eukaryotic cell line. Impact of combination Fe16 + AMP on chosen genes of S. aureus were performed by Quantitative Real-Time PCR.

RESULTS

The MIC of Fe16 + AMP was significantly lower than that of AMP and Fe16 alone. Furthermore, the infrared spectroscopy revealed the change in the ζ-potential of Fe16 + AMP. We demonstrated the ability of Fe16 + AMP to disrupt the bacterial membrane of S. aureus and that likely allowed for better absorption of AMP. In addition, the change in gene expression of bacterial efflux pumps at the sub-inhibitory concentration of AMP suggests an insufficient import of iron into the bacterial cell. At the same time, Fe16 + AMP did not have any cytotoxic effects on keratinocytes.

CONCLUSIONS

Combined Fe16 + AMP therapy demonstrated significant synergistic and antimicrobial effects against S. aureus. This study supports the potential of combination therapy and further research.

Synthesis and characterization of TiO2 nanoparticles combined with geraniol and their synergistic antibacterial activity

BACKGROUND

The emergence of antibiotic resistance in pathogenic bacteria has become a global threat, encouraging the adoption of efficient and effective alternatives to conventional antibiotics and promoting their use as replacements. Titanium dioxide nanoparticles (TiO₂ NPs) have been reported to exhibit antibacterial properties. In this study, we synthesized and characterized TiO₂ NPs in anatase and rutile forms with surface modification by geraniol (GER).

RESULTS

The crystallinity and morphology of modified TiO₂ NPs were analyzed by UV/Vis spectrophotometry, X-ray powder diffraction (XRD), and scanning electron microscopy (SEM) with elemental mapping (EDS). The antimicrobial activity of TiO₂ NPs with geraniol was assessed against Staphylococcus aureus, methicillin-resistant Staphylococcus aureus (MRSA), and Escherichia coli. The minimum inhibitory concentration (MIC) values of modified NPs ranged from 0.25 to 1.0 mg/ml against all bacterial strains, and the live dead assay and fractional inhibitory concentration (FIC) supported the antibacterial properties of TiO₂ NPs with GER. Moreover, TiO₂ NPs with GER also showed a significant decrease in the biofilm thickness of MRSA.

CONCLUSIONS

Our results suggest that TiO₂ NPs with GER offer a promising alternative to antibiotics, particularly for controlling antibiotic-resistant strains. The surface modification of TiO₂ NPs by geraniol resulted in enhanced antibacterial properties against multiple bacterial strains, including antibiotic-resistant MRSA. The potential applications of modified TiO₂ NPs in the biomedical and environmental fields warrant further investigation.

Genetic risk of depression is different in subgroups of dietary ratio of tryptophan to large neutral amino acids

Manipulation of intake of serotonin precursor tryptophan has been exploited to rapidly induce and alleviate depression symptoms. While studies show that this latter effect is dependent on genetic vulnerability to depression, the effect of habitual tryptophan intake in the context of predisposing genetic factors has not been explored. Our aim was to investigate the effect of habitual tryptophan intake on mood symptoms and to determine the effect of risk variants on depression in those with high and low tryptophan intake in the whole genome and specifically in serotonin and kynurenine pathways. 63,277 individuals in the UK Biobank with data on depressive symptoms and tryptophan intake were included. We compared two subpopulations defined by their habitual diet of a low versus a high ratio of tryptophan to other large amino acids (TLR). A modest protective effect of high dietary TLR against depression was found. NPBWR1 among serotonin genes and POLI in kynurenine pathway genes were significantly associated with depression in the low but not in the high TLR group. Pathway-level analyses identified significant associations for both serotonin and kynurenine pathways only in the low TLR group. In addition, significant association was found in the low TLR group between depressive symptoms and biological process related to adult neurogenesis. Our findings demonstrate a markedly distinct genetic risk profile for depression in groups with low and high dietary TLR, with association with serotonin and kynurenine pathway variants only in case of habitual food intake leading to low TLR. Our results confirm the relevance of the serotonin hypothesis in understanding the neurobiological background of depression and highlight the importance of understanding its differential role in the context of environmental variables such as complexity of diet in influencing mental health, pointing towards emerging possibilities of personalised prevention and intervention in mood disorders in those who are genetically vulnerable.

Detailed insight into chromium species released from failed CoCrMo implants: Ex vivo periprosthetic tissues study

This unique study provides information on Cr species and their distribution in periprosthetic tissues of patients with metal-on-polyethylene joint implants. Co-Cr-Mo alloy has been widely used in joint replacement and represents a source of metal derived species. In the case of chromium, previous studies on periprosthetic tissues revealed mainly Cr(III) distribution, whereas the potential release of carcinogenic Cr(VI) species has been still a subject of debate. Here, an analytical approach utilizing speciation and fractionation was developed to analyze periprosthetic tissue samples collected from wide range of patients with failed total hip or knee replacements. The results reveal that Cr(III) is mainly released in the form of insoluble CrPO₄ and Cr₂ O₃ particles. The highest Cr contents were found in periprosthetic tissues of patients suffering from aseptic loosening and having more Cr-based implants in the body. Cr species penetrated tissue layers, but their levels decreased with the distance from an implant. The detailed speciation/fractionation study carried out using the set of consecutive periprosthetic tissues of a patient with extensive metallosis showed the presence of trace amounts of free Cr(III), nanoparticles, and metal-protein complexes, but the majority of Cr still occurred in CrPO₄ form. Carcinogenic Cr(VI) species were not detected. Up to date, there is no published human tissue study focused on the detailed speciation of both soluble and insoluble Cr-based species in the context of failing total hip and knee replacements.

Body mass index and depressive rumination are positively associated with each other only in case of GG genotype of catenin alpha 2 gene rs13412541 variant

Introduction

Catenin alpha 2 gene (CTNNA2) is important in the stability of hippocampal synapses and also in brain development. Our recent paper (Eszlari et al, Pharmaceuticals 2021, 14, 850) has demonstrated that rumination on sad mood mediates the association of CTNNA2 only towards psychiatric symptoms, but not towards cardiovascular risk phenotypes.

Objectives

Our present aim was to test the moderating role of rumination and its two subtypes, brooding and reflection, in genetic associations between CTNNA2 and the same cardiovascular risk phenotypes.

Methods

633 unrelated subjects from the Budakalasz Health Examination Survey with non-missing phenotypic data, and 160 single-nucleotide CTNNA2 variants remaining after quality control, were included. Linear regression models were run in Plink 1.9 for separate outcomes of body mass index (BMI), and Framingham risk scores for cardiovascular disease, coronary heart disease, myocardial infarction, and stroke. With each variant, predictors were the variant, rumination or its subtype, the variant x rumination interaction, sex, age, and the top ten principal components of the genome. 100,000 label-swapping max(T) permutation was applied for the interaction term within each analysis.

Results

While no significant interaction term survived the familywise permutation, two trends emerged. Namely, BMI seems to have positive association with rumination and its maladaptive brooding subtype only in case of GG genotype of rs13412541, otherwise no association can be detected.

Conclusions

Although replication is needed in larger samples, the relationship between rumination and BMI, conditional on CTNNA2 genotype, can be important in atypical depression, thus may contribute to stratification of depressed patients.

Disclosure

The study was supported by the New National Excellence Program of the Ministry for Innovation and Technology from the source of the National Research, Development and Innovation Fund (ÚNKP-21-4-II-SE-1); and by 2019-2.1.7-ERA-NET-2020-00005.

The Genetic Underpinnings of Affective Temperaments: Identifying Novel Risk Variants with a Whole Genome Analytical Approach

One reason behind the failure to understand the neurobiological background of affective disorders and develop more effective treatments is their heterogeneity warranting identification of clinically meaningful endophenotypes. Affective temperaments, considered subclinical manifestations and pathoplastic contributors of affective illnesses may constitute such endophenotypes. 775 general population subjects were phenotyped for affective temperaments using TEMPS-A, and genotyped using Illumina’s CoreExom PsychChip yielding 573141 variants. Primary SNP-based association tests were calculated using linear regression models assuming an additive genetic effect with the first 10 calculated principal components, gender, age, and other affective temperaments as covariates; a Bonferroni-corrected genome-wide significance threshold set at p≤5.0×10⁻⁸, and suggestive significance threshold set at p≤1.0x10^-5. SNP-level relevances were aggregated to gene-level with the PEGASUS method, with a Bonferroni-corrected significance level set at 2.0×10^-6, and suggestive significance thrshold set at p≤4.0×10^-4. Functional effects of most significant SNPs as reported in public open databases based on expression quantitative trait loci (eQTL) and 3D-chromatin interactions were explored using FUMA v1.3.5. In SNP-based tests a novel genome-wide significant variant, rs3798978 (p=4.44x10^-8) and several other suggestively significant SNPs in ADGRB3 were found for anxious temperament along with suggestively significant SNPs for the other four affective temperaments. In gene-based tests suggestively significant findings emerged for all five temperaments. Functional analysis suggested that several of the identified variants influence gene-expression levels or participate in chromatin interactions in brain areas implicated in affective disorders. In the next step these findings should be investigated in patient samples, and in other models of affective disorders and related phenotypes.

Biology of Perseverative Negative Thinking: The Role of Timing and Folate Intake

Past-oriented rumination and future-oriented worry are two aspects of perseverative negative thinking related to the neuroticism endophenotype and associated with depression and anxiety. Our present aim was to investigate the genomic background of these two aspects of perseverative negative thinking within separate groups of individuals with suboptimal versus optimal folate intake. We conducted a genome-wide association study in the UK Biobank database (n = 72,621) on the "rumination" and "worry" items of the Eysenck Personality Inventory Neuroticism scale in these separate groups. Optimal folate intake was related to lower worry, but unrelated to rumination. In contrast, genetic associations for worry did not implicate specific biological processes, while past-oriented rumination had a more specific genetic background, emphasizing its endophenotypic nature. Furthermore, biological pathways leading to rumination appeared to differ according to folate intake: purinergic signaling and circadian regulator gene ARNTL emerged in the whole sample, blastocyst development, DNA replication, and C-C chemokines in the suboptimal folate group, and prostaglandin response and K+ channel subunit gene KCNH3 in the optimal folate group. Our results point to possible benefits of folate in anxiety disorders, and to the importance of simultaneously taking into account genetic and environmental factors to determine personalized intervention in polygenic and multifactorial disorders.

Catenin Alpha 2 May Be a Biomarker or Potential Drug Target in Psychiatric Disorders with Perseverative Negative Thinking

AlphaN-catenin gene CTNNA2 has been implicated in intrauterine brain development, as well as in several psychiatric disorders and cardiovascular diseases. Our present aim was to investigate CTNNA2 gene-wide associations of single-nucleotide polymorphisms (SNPs) with psychiatric and cardiovascular risk factors to test the potential mediating role of rumination, a perseverative negative thinking phenotype in these associations. Linear mixed regression models were run by FaST-LMM within a sample of 795 individuals from the Budakalasz Health Examination Survey. The psychiatric outcome variables were rumination and its subtypes, and ten Brief Symptom Inventory (BSI) scores including, e.g., obsessive-compulsive, depression, anxiety, hostility, phobic anxiety, and paranoid ideation. Cardiovascular outcome variables were BMI and the Framingham risk scores for cardiovascular disease, coronary heart disease, myocardial infarction, and stroke. We found nominally significant CTNNA2 associations for every phenotype. Rumination totally mediated the associations of CTNNA2 rs17019243 with eight out of ten BSI scores, but none with Framingham scores or BMI. Our results suggest that CTNNA2 genetics may serve as biomarkers, and increasing the expression or function of CTNNA2 protein may be a potential new therapeutic approach in psychiatric disorders with perseverative negative thinking including, e.g., depression. Generally, an antiruminative agent could be a transdiagnostic and preventive psychopharmacon.

Morphological remodeling of the intramural coronary resistance artery network geometry in chronically Angiotensin II infused hypertensive female rats

Segmental remodeling of resistance arteries, inhibition of angiogenetic processes, their rarefaction by AngiotensinII and hypertension are accepted facts. Less is known about alterations in resistance artery network geometry potentially induced by them. Female rats were infused with 100 ng/kg/min AngiotensinII with osmotic minipumps for four weeks that raised mean arterial blood pressure from 98 ± 3 to 125 ± 7 mmHg. Geometry of the left coronary artery system was studied on plastic casts and on in situ microsurgically prepared, saline infused video-microscoped networks (n = 13 and 11 controls and hypertensives, respectively). Parallel running branches, broken course of larger branches, multiple branchings and branch crossings have been identified (13 and 74 such deformities, in control and hypertensive networks, respectively, p < 0.01). Bifurcation angles increased with increasing asymmetry of daughter branches but not in hypertensives. Dividing the whole network (theoretically) into several hundreds of 50μm long ring units, ring frequency peaked at 200μm diameter in normal networks. This peak diminished and was replaced by a peak at 300μm in hypertensives (p < 0.01). In controls, diameter of vascular units decreased at a fairly even rate with flow distance from the orifice. The 350, 200, 150μm diameter units were found with highest frequencies at flow distances around 2.5, 5.5 and 7.5mm, respectively. This regular pattern disintegrated in hypertensives. Higher blood flow routes were needed to cover the same distance from the orifice (p < 0.01). Shrinkage and diminishment of many parallel connected 200μm segments, concomitant enlargement of many larger segments accompanied with morphological deformities can be expected to contribute to elevated vascular resistance.

Health and Aging: Unifying Concepts, Scores, Biomarkers and Pathways

Despite increasing research efforts, there is a lack of consensus on defining aging or health. To understand the underlying processes, and to foster the development of targeted interventions towards increasing one's health, there is an urgent need to find a broadly acceptable and useful definition of health, based on a list of (molecular) features; to operationalize features of health so that it can be measured; to identify predictive biomarkers and (molecular) pathways of health; and to suggest interventions, such as nutrition and exercise, targeted at putative causal pathways and processes. Based on a survey of the literature, we propose to define health as a state of an individual characterized by the core features of physiological, cognitive, physical and reproductive function, and a lack of disease. We further define aging as the aggregate of all processes in an individual that reduce its wellbeing, that is, its health or survival or both. We define biomarkers of health by their attribute of predicting future health better than chronological age. We define healthspan pathways as molecular features of health that relate to each other by belonging to the same molecular pathway. Our conceptual framework may integrate diverse operationalizations of health and guide precision prevention efforts.

The UKB envirome of depression: from interactions to synergistic effects

Major depressive disorder is a result of the complex interplay between a large number of environmental and genetic factors but the comprehensive analysis of contributing environmental factors is still an open challenge. The primary aim of this work was to create a Bayesian dependency map of environmental factors of depression, including life stress, social and lifestyle factors, using the UK Biobank data to determine direct dependencies and to characterize mediating or interacting effects of other mental health, metabolic or pain conditions. As a complementary approach, we also investigated the non-linear, synergistic multi-factorial risk of the UKB envirome on depression using deep neural network architectures. Our results showed that a surprisingly small number of core factors mediate the effects of the envirome on lifetime depression: neuroticism, current depressive symptoms, parental depression, body fat, while life stress and household income have weak direct effects. Current depressive symptom showed strong or moderate direct relationships with life stress, pain conditions, falls, age, insomnia, weight change, satisfaction, confiding in someone, exercise, sports and Townsend index. In conclusion, the majority of envirome exerts their effects in a dynamic network via transitive, interactive and synergistic relationships explaining why environmental effects may be obscured in studies which consider them individually.

Genome-wide association analysis reveals KCTD12 and miR-383-binding genes in the background of rumination

Ruminative response style is a passive and repetitive way of responding to stress, associated with several disorders. Although twin and candidate gene studies have proven the genetic underpinnings of rumination, no genome-wide association study (GWAS) has been conducted yet. We performed a GWAS on ruminative response style and its two subtypes, brooding and reflection, among 1758 European adults recruited in the general population of Budapest, Hungary, and Manchester, United Kingdom. We evaluated single-nucleotide polymorphism (SNP)-based, gene-based and gene set-based tests, together with inferences on genes regulated by our most significant SNPs. While no genome-wide significant hit emerged at the SNP level, the association of rumination survived correction for multiple testing with KCTD12 at the gene level, and with the set of genes binding miR-383 at the gene set level. SNP-level results were concordant between the Budapest and Manchester subsamples for all three rumination phenotypes. SNP-level results and their links to brain expression levels based on external databases supported the role of KCTD12, SRGAP3, and SETD5 in rumination, CDH12 in brooding, and DPYSL5, MAPRE3, KCNK3, ATXN7L3B, and TPH2 in reflection, among others. The relatively low sample size is a limitation of our study. Results of the first GWAS on rumination identified genes previously implicated in psychiatric disorders underscoring the transdiagnostic nature of rumination, and pointed to the possible role of the dorsolateral prefrontal cortex, hippocampus, and cerebellum in this cognitive process.

Ion-pair complexes of Schiff base Fe(iii) cations and complex anions

We report on the crystal structure and magnetic properties of four new ion-pair complexes.

Genetic variants in major depressive disorder: From pathophysiology to therapy

In spite of promising preclinical results there is a decreasing number of new registered medications in major depression. The main reason behind this fact is the lack of confirmation in clinical studies for the assumed, and in animals confirmed, therapeutic results. This suggests low predictive value of animal studies for central nervous system disorders. One solution for identifying new possible targets is the application of genetics and genomics, which may pinpoint new targets based on the effect of genetic variants in humans. The present review summarizes such research focusing on depression and its therapy. The inconsistency between most genetic studies in depression suggests, first of all, a significant role of environmental stress. Furthermore, effect of individual genes and polymorphisms is weak, therefore gene x gene interactions or complete biochemical pathways should be analyzed. Even genes encoding target proteins of currently used antidepressants remain non-significant in genome-wide case control investigations suggesting no main effect in depression, but rather an interaction with stress. The few significant genes in GWASs are related to neurogenesis, neuronal synapse, cell contact and DNA transcription and as being nonspecific for depression are difficult to harvest pharmacologically. Most candidate genes in replicable gene x environment interactions, on the other hand, are connected to the regulation of stress and the HPA axis and thus could serve as drug targets for depression subgroups characterized by stress-sensitivity and anxiety while other risk polymorphisms such as those related to prominent cognitive symptoms in depression may help to identify additional subgroups and their distinct treatment. Until these new targets find their way into therapy, the optimization of current medications can be approached by pharmacogenomics, where metabolizing enzyme polymorphisms remain prominent determinants of therapeutic success.

Probabilistic Graphical Models for Computational Biomedicine

Background: As genomics becomes increasingly relevant to medicine, medical informatics and bioinformatics are gradually converging into a larger field that we call computational biomedicine.

Objectives: Developing a computational framework that is common to the different disciplines that compose computational biomedicine will be a major enabler of the further development and integration of this research domain.

Methods: Probabilistic graphical models such as Hidden Markov Models, belief networks, and missing-data models together with computational methods such as dynamic programming, Expectation-Maximization, data-augmentation Gibbs sampling, and the Metropolis-Hastings algorithm provide the tools for an integrated probabilistic approach to computational biomedicine.

Results and Conclusions: We show how graphical models have already found a broad application in different fields composing computational biomedicine. We also indicate several challenges that lie at the interface between medical informatics, statistical genomics, and bioinformatics. We also argue that graphical models offer a unified framework making it possible to integrate in a statistically meaningful way multiple models ranging from the molecular level to cellular and to clinical levels. Because of their versatility and firm statistical underpinning, we assert that probabilistic graphical models can serve as the lingua franca for many computationally intensive approaches to biology and medicine. As such, graphical models should be a foundation of the curriculum of students in these fields. From such a foundation, students could then build towards specific computational methods in medical informatics, medical image analysis, statistical genetics, or bioinformatics while keeping the communication open between these areas.

Comorbidities in the diseasome are more apparent than real: What Bayesian filtering reveals about the comorbidities of depression

Comorbidity patterns have become a major source of information to explore shared mechanisms of pathogenesis between disorders. In hypothesis-free exploration of comorbid conditions, disease-disease networks are usually identified by pairwise methods. However, interpretation of the results is hindered by several confounders. In particular a very large number of pairwise associations can arise indirectly through other comorbidity associations and they increase exponentially with the increasing breadth of the investigated diseases. To investigate and filter this effect, we computed and compared pairwise approaches with a systems-based method, which constructs a sparse Bayesian direct multimorbidity map (BDMM) by systematically eliminating disease-mediated comorbidity relations. Additionally, focusing on depression-related parts of the BDMM, we evaluated correspondence with results from logistic regression, text-mining and molecular-level measures for comorbidities such as genetic overlap and the interactome-based association score. We used a subset of the UK Biobank Resource, a cross-sectional dataset including 247 diseases and 117,392 participants who filled out a detailed questionnaire about mental health. The sparse comorbidity map confirmed that depressed patients frequently suffer from both psychiatric and somatic comorbid disorders. Notably, anxiety and obesity show strong and direct relationships with depression. The BDMM identified further directly co-morbid somatic disorders, e.g. irritable bowel syndrome, fibromyalgia, or migraine. Using the subnetwork of depression and metabolic disorders for functional analysis, the interactome-based system-level score showed the best agreement with the sparse disease network. This indicates that these epidemiologically strong disease-disease relations have improved correspondence with expected molecular-level mechanisms. The substantially fewer number of comorbidity relations in the BDMM compared to pairwise methods implies that biologically meaningful comorbid relations may be less frequent than earlier pairwise methods suggested. The computed interactive comprehensive multimorbidity views over the diseasome are available on the web at Co=MorNet: bioinformatics.mit.bme.hu/UKBNetworks.

Variants in the CNR1 gene predispose to headache with nausea in the presence of life stress

One of the main effects of the endocannabinoid system in the brain is stress adaptation with presynaptic endocannabinoid receptor 1 (CB1 receptors) playing a major role. In the present study, we investigated whether the effect of the CB1 receptor coding CNR1 gene on migraine and its symptoms is conditional on life stress. In a cross-sectional European population (n = 2426), recruited from Manchester and Budapest, we used the ID-Migraine questionnaire for migraine screening, the Life Threatening Experiences questionnaire to measure recent negative life events (RLE), and covered the CNR1 gene with 11 SNPs. The main genetic effects and the CNR1 × RLE interaction with age and sex as covariates were tested. None of the SNPs showed main genetic effects on possible migraine or its symptoms, but 5 SNPs showed nominally significant interaction with RLE on headache with nausea using logistic regression models. The effect of rs806366 remained significant after correction for multiple testing and replicated in the subpopulations. This effect was independent from depression- and anxiety-related phenotypes. In addition, a Bayesian systems-based analysis demonstrated that in the development of headache with nausea all SNPs were more relevant with higher a posteriori probability in those who experienced recent life stress. In summary, the CNR1 gene in interaction with life stress increased the risk of headache with nausea suggesting a specific pathological mechanism to develop migraine, and indicating that a subgroup of migraine patients, who suffer from life stress triggered migraine with frequent nausea, may benefit from therapies that increase the endocannabinoid tone.

Muffin-like lanthanide complexes with an N5O2-donor macrocyclic ligand showing field-induced single-molecule magnet behaviour

Three mononuclear lanthanide complexes of a 2-pyridylmethyl pendant-armed 15-membered ligand {L} with general formula [Ln(L)(H₂O)(NO₃)](NO₃)₂ (Ln = Tb (1), Dy (2), and Er (3)) are reported.

Variability in the effect of 5-HTTLPR on depression in a large European population: the role of age, symptom profile, type and intensity of life stressors

Background

Although 5-HTTLPR has been shown to influence the risk of life stress-induced depression in the majority of studies, others have produced contradictory results, possibly due to weak effects and/or sample heterogeneity.

Methods

In the present study we investigated how age, type and intensity of life-stressors modulate the effect of 5-HTTLPR on depression and anxiety in a European population cohort of over 2300 subjects. Recent negative life events (RLE), childhood adversity (CHA), lifetime depression, Brief Symptoms Inventory (BSI) depression and anxiety scores were determined in each subject. Besides traditional statistical analysis we calculated Bayesian effect strength and relevance of 5-HTTLPR genotypes in specified models.

Results

The short (s) low expressing allele showed association with increased risk of depression related phenotypes, but all nominally significant effects would turn to non-significant after correction for multiple testing in the traditional analysis. Bayesian effect strength and relevance analysis, however, confirmed the role of 5-HTTLPR. Regarding current (BSI) and lifetime depression 5-HTTLPR-by-RLE interactions were confirmed. Main effect, with other words direct association, was supported with BSI anxiety. With more frequent RLE the prevalence or symptoms of depression increased in ss carriers. Although CHA failed to show an interaction with 5-HTTLPR, in young subjects CHA sensitized towards the depression promoting effect of even mild RLE. Furthermore, the direct association of anxiety with the s allele was driven by young (≤30) individuals.

Limitations

Our study is cross-sectional and applies self-report questionnaires.

Conclusions

Albeit 5-HTTLPR has only weak/moderate effects, the s allele is directly associated with anxiety and modulates development of depression in homogeneous subgroups.

In interaction with gender a common CYP3A4 polymorphism may influence the survival rate of chemotherapy for childhood acute lymphoblastic leukemia

CYP3A4 has an important role in the metabolisms of many drugs used in acute lymphoblastic leukemia (ALL) therapy; still, there are practically no publications about the role of CYP3A4 polymorphisms in ALL pharmacogenomics. We genotyped eight common single-nucleotide polymorphisms (SNPs) in the CYP3A4 and CYP3A5 genes in 511 children with ALL and investigated whether they influenced the survival of the patients. We involved additional 127 SNPs in 34 candidate genes and searched for interactions with respect to the survival rates. Significant association between the survival rates and the common rs2246709 SNP in the CYP3A4 gene was observed. The gender of the patients and the rs1076991 in the MTHFD1 gene strongly influenced this effect. We calculated new risk assessments involving the gender-rs2246709 interaction and showed that they significantly outperformed the earlier risk-group assessments at every time point. If this finding is confirmed in other populations, it can have a considerable prognostic significance.

Multivariate analysis of dopaminergic gene variants as risk factors of heroin dependence

BACKGROUND

Heroin dependence is a debilitating psychiatric disorder with complex inheritance. Since the dopaminergic system has a key role in rewarding mechanism of the brain, which is directly or indirectly targeted by most drugs of abuse, we focus on the effects and interactions among dopaminergic gene variants.

OBJECTIVE

To study the potential association between allelic variants of dopamine D2 receptor (DRD2), ANKK1 (ankyrin repeat and kinase domain containing 1), dopamine D4 receptor (DRD4), catechol-O-methyl transferase (COMT) and dopamine transporter (SLC6A3) genes and heroin dependence in Hungarian patients.

METHODS

303 heroin dependent subjects and 555 healthy controls were genotyped for 7 single nucleotide polymorphisms (SNPs) rs4680 of the COMT gene; rs1079597 and rs1800498 of the DRD2 gene; rs1800497 of the ANKK1 gene; rs1800955, rs936462 and rs747302 of the DRD4 gene. Four variable number of tandem repeats (VNTRs) were also genotyped: 120 bp duplication and 48 bp VNTR in exon 3 of DRD4 and 40 bp VNTR and intron 8 VNTR of SLC6A3. We also perform a multivariate analysis of associations using Bayesian networks in Bayesian multilevel analysis (BN-BMLA).

FINDINGS AND CONCLUSIONS

In single marker analysis the TaqIA (rs1800497) and TaqIB (rs1079597) variants were associated with heroin dependence. Moreover, -521 C/T SNP (rs1800955) of the DRD4 gene showed nominal association with a possible protective effect of the C allele. After applying the Bonferroni correction TaqIB was still significant suggesting that the minor (A) allele of the TaqIB SNP is a risk component in the genetic background of heroin dependence. The findings of the additional multiple marker analysis are consistent with the results of the single marker analysis, but this method was able to reveal an indirect effect of a promoter polymorphism (rs936462) of the DRD4 gene and this effect is mediated through the -521 C/T (rs1800955) polymorphism in the promoter.

Multi-aspect candidates for repositioning: data fusion methods using heterogeneous information sources

Drug repositioning, an innovative therapeutic application of an old drug, has received much attention as a particularly costeffective strategy in drug R&D Recent work has indicated that repositioning can be promoted by utilizing a wide range of information sources, including medicinal chemical, target, mechanism, main and side-effect-related information, and also bibliometric and taxonomical fingerprints, signatures and knowledge bases. This article describes the adaptation of a conceptually novel, more efficient approach for the identification of new possible therapeutic applications of approved drugs and drug candidates, based on a kernel-based data fusion method. This strategy includes (1) the potentially multiple representation of information sources, (2) the automated weighting and statistically optimal combination of information sources, and (3) the automated weighting of parts of the query compounds. The performance was systematically evaluated by using Anatomical Therapeutic Chemical Classification System classes in a cross-validation framework. The results confirmed that kernel-based data fusion can integrate heterogeneous information sources significantly better than standard rank-based fusion can, and this method provides a unique solution for repositioning; it can also be utilized for de novo drug discovery. The advantages of kernel-based data fusion are illustrated with examples and open problems that are particularly relevant for pharmaceutical applications.

Beyond structural equation modeling: model properties and effect size from a Bayesian viewpoint. An example of complex phenotype-genotype associations in depression

Despite the rapid evolution of measurement technologies in biomedicine and genetics, most of the recent studies aiming to explore the genetic background of multifactorial diseases were only moderately successful. One of the causes of this phenomenon is that the bottleneck of genetic research is no longer the measurement process related to various laboratory technologies, but rather the analysis and interpretation of results. The commonly applied univariate methods are inadequate for exploring complex dependency patterns of multifactorial diseases which includes nearly all common diseases, such as depression, hypertension, and asthma. A comprehensive investigation requires multivariate modeling methods that enable the analysis of interactions between factors, and allow a more detailed interpretation of studies measuring complex phenotype descriptors. In this paper we discuss various aspects of multivariate modeling through a case study analyzing the effect of the single nucleotide polymorphism rs6295 in the HTR1A gene on depression and impulsivity. We overview basic concepts related to multivariate modeling and compare the properties of two investigated modeling techniques: Structural Equation Modeling and Bayesian network based learning algorithms. The resulting models demonstrate the advantages of the Bayesian approach in terms of model properties and effect size as it allows coherent handling of the weakly significant effect of rs6295. Results also confirm the mediating role of impulsivity between the SNP rs6295 of HTR1A and depression.

Pharmacological reactivity of resistance vessels in a rat PCOS model - vascular effects of parallel vitamin D₃ treatment

The aim of this study was to clarify the effects of dihydrotestosterone (DHT)-induced polycystic ovary syndrome (PCOS) on pharmacological reactivity of a resistance vessel in a rat model and the possible modulatory role of 1,25-(OH)₂-cholecalciferol (vitamin D₃). The PCOS model was induced in adolescent female Wistar rats by a 10-week DHT treatment. Norepinephrine induced contractility and acetylcholine relaxation were tested in arterioles by pressure arteriography in control as well as DHT- and DHT plus vitamin D₃-treated (DHT+D3) animals. Decreased vasoconstriction and dilatation were detected after DHT treatment. Concomitant vitamin D₃ treatment increased the contractile response and resulted in more relaxed vessels. Endothelial dilation tested with acetylcholine was lower after DHT treatment, this effect was not depend on vitamin D₃ supplementation. In conclusion, hyperandrogenic state resulted in reduced endothelium- and smooth muscle-dependent vasorelaxation and constriction with a complete loss of nitric oxide (NO)-dependent relaxation compared with controls. These alterations caused by chronic DHT treatment were partially reversed by concomitant vitamin D₃ administration.

Arteriolar biomechanics in a rat polycystic ovary syndrome model - effects of parallel vitamin D3 treatment

To clarify the effects of dihydrotestosterone (DHT)-induced polycystic ovary syndrome (PCOS) on arteriolar biomechanics in a rat model and the possible modulatory role of vitamin D3.

METHODS AND RESULTS

The PCOS model was induced in female Wistar rats by ten-weeks DHT treatment. Arteriolar biomechanics was tested in arterioles by pressure arteriography in control as well as DHT- and DHT with vitamin D3-treated animals in contracted and passive conditions. Increased wall stress and distensibility as well as increased vascular lumen were detected after DHT treatment. Concomitant vitamin D3 treatment lowered the mechanical load of the arterioles and restored the vascular diameter.

CONCLUSION

The hyperandrogenic state resulted in more rigid, less flexible arteriolar walls with increased vascular lumen compared with controls. DHT treatment caused eutrophic remodelling of gracilis arteriole. These prehypertensive alterations caused by chronic DHT treatment were mostly reversed by concomitant vitamin D3 administration.

Additive effects of serotonergic and dopaminergic polymorphisms on trait impulsivity

Twin studies suggest 45% heritability of trait impulsivity. Results from candidate gene studies to date are contradictory; impulsivity phenotypes were measured by different behavioral and questionnaire methods related either to the dopaminergic or to the serotonergic system. Here we report an association study of both dopaminergic (COMT rs4680, DRD4 48 bp VNTR, DRD2/ANKK1 rs1800497) and serotonergic (HTR1A rs6925, HTR1B rs13212041, SLC6A4 5-HTTLPR) gene polymorphisms and trait impulsivity assessed with the Barratt Impulsiveness Scale (BIS-11) in a sample of 687 Caucasian young adults. Results showed lower impulsivity in the presence of the DRD4 7-repeat (P = 0.006) and the HTR1B rs13212041 alleles (P = 0.003). These findings stayed significant after Bonferroni correction. A multivariate analysis using Bayesian networks confirmed independent effects of these two polymorphisms and provided a coherent characterization of the system of dependencies with respect to the impulsivity construct as well as its subscales. These results clearly suggest an additive effect of dopaminergic and serotonergic polymorphisms on trait impulsivity.

Arteriolar insulin resistance in a rat model of polycystic ovary syndrome

OBJECTIVE

To investigate the vascular dysfunction caused by insulin resistance in polycystic ovary syndrome (PCOS) and the effectiveness of vitamin D in an animal model.

DESIGN

Controlled experimental animal study.

SETTING

Animal laboratory at a university research institute.

ANIMAL(S)

Thirty female Wistar rats.

INTERVENTION(S)

Rats were divided into groups at age 21-28 weeks. Twenty of them were subjected to dihydrotestosterone (DHT) treatment (83 μg/d); ten of them also received parallel vitamin D treatment (120 ng/100 g/wk). Oral glucose tolerance tests with insulin level measurements were performed. Gracilis arterioles were tested for their contractility as well as their nitric oxide (NO)-dependent and insulin-induced dilation using pressure arteriography.

MAIN OUTCOME MEASURE(S)

Several physiologic parameters, glucose metabolism, and pressure arteriography.

RESULT(S)

DHT treatment increased the passive diameter of resistance arterioles, lowered norepinephrine-induced contraction (30.1 ± 4.7% vs. 8.7 ± 3.6%) and reduced acetylcholine-induced (122.0 ± 2.9% vs. 48.0 ± 1.4%) and insulin-induced (at 30 mU/mL: 21.7 ± 5.3 vs. 9.8 ± 5.6%) dilation. Vitamin D treatment restored insulin relaxation and norepinephrine-induced contractility; in contrast, it failed to alter NO-dependent relaxation.

CONCLUSION(S)

In DHT-treated rats, in addition to metabolically proven insulin resistance, decreased insulin-induced vasorelaxation was observed and was improved by vitamin D treatment without affecting NO-dependent relaxation. The reduction in insulin-induced dilation of arterioles is an important as yet undescribed pathway of vascular damage in PCOS and might explain the clinical effectiveness of vitamin D treatment.

Using literature and data to learn Bayesian networks as clinical models of ovarian tumors

Thanks to its increasing availability, electronic literature has become a potential source of information for the development of complex Bayesian networks (BN), when human expertise is missing or data is scarce or contains much noise. This opportunity raises the question of how to integrate information from free-text resources with statistical data in learning Bayesian networks. Firstly, we report on the collection of prior information resources in the ovarian cancer domain, which includes "kernel" annotations of the domain variables. We introduce methods based on the annotations and literature to derive informative pairwise dependency measures, which are derived from the statistical cooccurrence of the names of the variables, from the similarity of the "kernel" descriptions of the variables and from a combined method. We perform wide-scale evaluation of these text-based dependency scores against an expert reference and against data scores (the mutual information (MI) and a Bayesian score). Next, we transform the text-based dependency measures into informative text-based priors for Bayesian network structures. Finally, we report the benefit of such informative text-based priors on the performance of a Bayesian network for the classification of ovarian tumors from clinical data.

Bayesian applications of belief networks and multilayer perceptrons for ovarian tumor classification with rejection

Incorporating prior knowledge into black-box classifiers is still much of an open problem. We propose a hybrid Bayesian methodology that consists in encoding prior knowledge in the form of a (Bayesian) belief network and then using this knowledge to estimate an informative prior for a black-box model (e.g. a multilayer perceptron). Two technical approaches are proposed for the transformation of the belief network into an informative prior. The first one consists in generating samples according to the most probable parameterization of the Bayesian belief network and using them as virtual data together with the real data in the Bayesian learning of a multilayer perceptron. The second approach consists in transforming probability distributions over belief network parameters into distributions over multilayer perceptron parameters. The essential attribute of the hybrid methodology is that it combines prior knowledge and statistical data efficiently when prior knowledge is available and the sample is of small or medium size. Additionally, we describe how the Bayesian approach can provide uncertainty information about the predictions (e.g. for classification with rejection). We demonstrate these techniques on the medical task of predicting the malignancy of ovarian masses and summarize the practical advantages of the Bayesian approach. We compare the learning curves for the hybrid methodology with those of several belief networks and multilayer perceptrons. Furthermore, we report the performance of Bayesian belief networks when they are allowed to exclude hard cases based on various measures of prediction uncertainty.

Evaluation of the vector space representation in text-based gene clustering

Thanks to its increasing availability, electronic literature can now be a major source of information when developing complex statistical models where data is scarce or contains much noise. This raises the question of how to deeply integrate information from domain literature with experimental data. Evaluating what kind of statistical text representations can integrate literature knowledge in clustering still remains an unsufficiently explored topic. In this work we discuss how the bag-of-words representation can be used successfully to represent genetic annotation and free-text information coming from different databases. We demonstrate the effect of various weighting schemes and information sources in a functional clustering setup. As a quantitative evaluation, we contrast for different parameter settings the functional groupings obtained from text with those obtained from expert assessments and link each of the results to a biological discussion.

Probabilistic graphical models for computational biomedicine

BACKGROUND

As genomics becomes increasingly relevant to medicine, medical informatics and bioinformatics are gradually converging into a larger field that we call computational biomedicine.

OBJECTIVES

Developing a computational framework that is common to the different disciplines that compose computational biomedicine will be a major enabler of the further development and integration of this research domain.

METHODS

Probabilistic graphical models such as Hidden Markov Models, belief networks, and missing-data models together with computational methods such as dynamic programming, Expectation-Maximization, data-augmentation Gibbs sampling, and the Metropolis-Hastings algorithm provide the tools for an integrated probabilistic approach to computational biomedicine.

RESULTS AND CONCLUSIONS

We show how graphical models have already found a broad application in different fields composing computational biomedicine. We also indicate several challenges that lie at the interface between medical informatics, statistical genomics, and bioinformatics. We also argue that graphical models offer a unified framework making it possible to integrate in a statistically meaningful way multiple models ranging from the molecular level to cellular and to clinical levels. Because of their versatility and firm statistical underpinning, we assert that probabilistic graphical models can serve as the lingua franca for many computationally intensive approaches to biology and medicine. As such, graphical models should be a foundation of the curriculum of students in these fields. From such a foundation, students could then build towards specific computational methods in medical informatics, medical image analysis, statistical genetics, or bioinformatics while keeping the communication open between these areas.