Pharmacogenomic and clinical data link non-pharmacokinetic metabolic dysregulation to drug side effect pathogenesis

MSDSE: Predicting drug-side effects based on multi-scale features and deep multi-structure neural network

Unexpected side effects may accompany the research stage and post-marketing of drugs. These accidents lead to drug development failure and even endanger patients' health. Thus, it is essential to recognize the unknown drug-side effects. Most existing methods in silico find the answer from the association network or similarity network of drugs while ignoring the drug-intrinsic attributes. The limitation is that they can only handle drugs in the maturation stage. To be suitable for early drug-side effect screening, we conceive a multi-structural deep learning framework, MSDSE, which synthetically considers the multi-scale features derived from the drug. MSDSE can jointly learn SMILES sequence-based word embedding, substructure-based molecular fingerprint, and chemical structure-based graph embedding. In the preprocessing stage of MSDSE, we project all features to the abstract space with the same dimension. MSDSE builds a bi-level channel strategy, including a convolutional neural network module with an Inception structure and a multi-head Self-Attention module, to learn and integrate multi-modal features from local to global perspectives. Finally, MSDSE regards the prediction of drug-side effects as pair-wise learning and outputs the pair-wise probability of drug-side effects through the inner product operation. MSDSE is evaluated and analyzed on benchmark datasets and performs optimally compared to other baseline models. We also set up the ablation study to explain the rationality of the feature approach and model structure. Moreover, we select model partial prediction results for the case study to reveal actual capability. The original data are available at http://github.com/yuliyi/MSDSE.

QuTIE: quantum optimization for target identification by enzymes

Summary

Target identification by enzymes (TIE) problem aims to identify the set of enzymes in a given metabolic network, such that their inhibition eliminates a given set of target compounds associated with a disease while incurring minimum damage to the rest of the compounds. This is a NP-hard problem, and thus optimal solutions using classical computers fail to scale to large metabolic networks. In this article, we develop the first quantum optimization solution, called QuTIE (quantum optimization for target identification by enzymes), to this NP-hard problem. We do that by developing an equivalent formulation of the TIE problem in quadratic unconstrained binary optimization form. We then map it to a logical graph, and embed the logical graph on a quantum hardware graph. Our experimental results on 27 metabolic networks from Escherichia coli, Homo sapiens, and Mus musculus show that QuTIE yields solutions that are optimal or almost optimal. Our experiments also demonstrate that QuTIE can successfully identify enzyme targets already verified in wet-lab experiments for 14 major disease classes.

Availability and implementation

Code and sample data are available at: https://github.com/ngominhhoang/Quantum-Target-Identification-by-Enzymes.

What are side effects?

Side effects are ubiquitous in medicine and they often play a role in treatment decisions for patients and clinicians alike. Philosophers and health researchers often use side effects to illustrate issues with contemporary medical research and practice. However, technical definitions of 'side effect' differ among health authorities. Thus, determining the side effects of an intervention can differ depending on whose definition we assume. Here I review some of the common definitions of side effect and highlight their issues. In response, I offer an account of side effects as jointly (i) unintended and (ii) effects due to the causal capacities or invariances of an intervention. I discuss (i) by examining the intentions or reasons behind therapeutic interventions, and I discuss (ii) by appealing to a manipulationist model of causation. The analysis here highlights that side effects are conceptually distinct from related outcomes like adverse events, adverse drug reactions, and placebo effects. The analysis also allows for reflection on the utility of 'side effect' as a technical term in medical research and practice.

DeepSide: A Deep Learning Approach for Drug Side Effect Prediction

Drug failures due to unforeseen adverse effects at clinical trials pose health risks for the participants and lead to substantial financial losses. Side effect prediction algorithms have the potential to guide the drug design process. LINCS L1000 dataset provides a vast resource of cell line gene expression data perturbed by different drugs and creates a knowledge base for context specific features. The state-of-the-art approach that aims at using context specific information relies on only the high-quality experiments in LINCS L1000 and discards a large portion of the experiments. In this study, our goal is to boost the prediction performance by utilizing this data to its full extent. We experiment with 5 deep learning architectures. We find that a multi-modal architecture produces the best predictive performance among multi-layer perceptron-based architectures when drug chemical structure (CS), and the full set of drug perturbed gene expression profiles (GEX) are used as modalities. Overall, we observe that the CS is more informative than the GEX. A convolutional neural network-based model that uses only SMILES string representation of the drugs achieves the best results and provides 13.0% macro-AUC and 3.1% micro-AUC improvements over the state-of-the-art. We also show that the model is able to predict side effect-drug pairs that are reported in the literature but was missing in the ground truth side effect dataset. DeepSide is available at http://github.com/OnurUner/DeepSide.

Genome-scale integration of transcriptome and metabolome unveils squalene synthase and dihydrofolate reductase as targets against AML cells resistant to chemotherapy

The development of resistance to chemotherapeutic agents, such as Doxorubicin (DOX) and cytarabine (AraC), is one of the greatest challenges to the successful treatment of Acute Myeloid Leukemia (AML). Such acquisition is often underlined by a metabolic reprogramming that can provide a therapeutic opportunity, as it can lead to the emergence of vulnerabilities and dependencies to be exploited as targets against the resistant cells. In this regard, genome-scale metabolic models (GSMMs) have emerged as powerful tools to integrate multiple layers of data to build cancer-specific models and identify putative metabolic vulnerabilities. Here, we use genome-scale metabolic modelling to reconstruct a GSMM of the THP1 AML cell line and two derivative cell lines, one with acquired resistance to AraC and the second with acquired resistance to DOX. We also explore how, adding to the transcriptomic layer, the metabolomic layer enhances the selectivity of the resulting condition specific reconstructions. The resulting models enabled us to identify and experimentally validate that drug-resistant THP1 cells are sensitive to the FDA-approved antifolate methotrexate. Moreover, we discovered and validated that the resistant cell lines could be selectively targeted by inhibiting squalene synthase, providing a new and promising strategy to directly inhibit cholesterol synthesis in AML drug resistant cells.

Identifying functional metabolic shifts in heart failure with the integration of omics data and a heart-specific, genome-scale model

In diseased states, the heart can shift to use different carbon substrates, measured through changes in uptake of metabolites by imaging methods or blood metabolomics. However, it is not known whether these measured changes are a result of transcriptional changes or external factors. Here, we explore transcriptional changes in late-stage heart failure using publicly available data integrated with a model of heart metabolism. First, we present a heart-specific genome-scale metabolic network reconstruction (GENRE), iCardio. Next, we demonstrate the utility of iCardio in interpreting heart failure gene expression data by identifying tasks inferred from differential expression (TIDEs), which represent metabolic functions associated with changes in gene expression. We identify decreased gene expression for nitric oxide (NO) and N-acetylneuraminic acid (Neu5Ac) synthesis as common metabolic markers of heart failure. The methods presented here for constructing a tissue-specific model and identifying TIDEs can be extended to multiple tissues and diseases of interest.

The Metano Modeling Toolbox MMTB: An Intuitive, Web-Based Toolbox Introduced by Two Use Cases

Genome-scale metabolic models are of high interest in a number of different research fields. Flux balance analysis (FBA) and other mathematical methods allow the prediction of the steady-state behavior of metabolic networks under different environmental conditions. However, many existing applications for flux optimizations do not provide a metabolite-centric view on fluxes. Metano is a standalone, open-source toolbox for the analysis and refinement of metabolic models. While flux distributions in metabolic networks are predominantly analyzed from a reaction-centric point of view, the Metano methods of split-ratio analysis and metabolite flux minimization also allow a metabolite-centric view on flux distributions. In addition, we present MMTB (Metano Modeling Toolbox), a web-based toolbox for metabolic modeling including a user-friendly interface to Metano methods. MMTB assists during bottom-up construction of metabolic models by integrating reaction and enzymatic annotation data from different databases. Furthermore, MMTB is especially designed for non-experienced users by providing an intuitive interface to the most commonly used modeling methods and offering novel visualizations. Additionally, MMTB allows users to upload their models, which can in turn be explored and analyzed by the community. We introduce MMTB by two use cases, involving a published model of Corynebacterium glutamicum and a newly created model of Phaeobacter inhibens.

A hexokinase isoenzyme switch in human liver cancer cells promotes lipogenesis and enhances innate immunity

During the cancerous transformation of normal hepatocytes into hepatocellular carcinoma (HCC), the enzyme catalyzing the first rate-limiting step of glycolysis, namely the glucokinase (GCK), is replaced by the higher affinity isoenzyme, hexokinase 2 (HK2). Here, we show that in HCC tumors the highest expression level of HK2 is inversely correlated to GCK expression, and is associated to poor prognosis for patient survival. To further explore functional consequences of the GCK-to-HK2 isoenzyme switch occurring during carcinogenesis, HK2 was knocked-out in the HCC cell line Huh7 and replaced by GCK, to generate the Huh7-GCK⁺/HK2⁻ cell line. HK2 knockdown and GCK expression rewired central carbon metabolism, stimulated mitochondrial respiration and restored essential metabolic functions of normal hepatocytes such as lipogenesis, VLDL secretion, glycogen storage. It also reactivated innate immune responses and sensitivity to natural killer cells, showing that consequences of the HK switch extend beyond metabolic reprogramming.

Many cancers fuel their rapid growth by replacing glucokinase with its higher affinity isoenzyme, hexokinase 2 (HK2), making HK2 an attractive drug target. In this study, Perrin-Cocon and Vidalain et al. use CRISPR/Cas-9 gene editing to reverse this enzymatic switch in human liver cancer cells, and find this restores innate immune function as well as reversing cancer-associated metabolic reprogramming.

FBA reveals guanylate kinase as a potential target for antiviral therapies against SARS-CoV-2

Motivation

The novel coronavirus (SARS-CoV-2) currently spreads worldwide, causing the disease COVID-19. The number of infections increases daily, without any approved antiviral therapy. The recently released viral nucleotide sequence enables the identification of therapeutic targets, e.g. by analyzing integrated human-virus metabolic models. Investigations of changed metabolic processes after virus infections and the effect of knock-outs on the host and the virus can reveal new potential targets.

Results

We generated an integrated host–virus genome-scale metabolic model of human alveolar macrophages and SARS-CoV-2. Analyses of stoichiometric and metabolic changes between uninfected and infected host cells using flux balance analysis (FBA) highlighted the different requirements of host and virus. Consequently, alterations in the metabolism can have different effects on host and virus, leading to potential antiviral targets. One of these potential targets is guanylate kinase (GK1). In FBA analyses, the knock-out of the GK1 decreased the growth of the virus to zero, while not affecting the host. As GK1 inhibitors are described in the literature, its potential therapeutic effect for SARS-CoV-2 infections needs to be verified in in-vitro experiments.

Availability and implementation

The computational model is accessible at https://identifiers.org/biomodels.db/MODEL2003020001.

Metabolic Network Reconstructions to Predict Drug Targets and Off-Target Effects

The drug development pipeline has stalled because of the difficulty in identifying new drug targets while minimizing off-target effects. Computational methods, such as the use of metabolic network reconstructions, may provide a cost-effective platform to test new hypotheses for drug targets and prevent off-target effects. Here, we summarize available methods to identify drug targets and off-target effects using either reaction-centric, gene-centric, or metabolite-centric approaches with genome-scale metabolic network reconstructions.

Opportunities for Artificial Intelligence in Advancing Precision Medicine

PURPOSE OF REVIEW

We critically evaluate the future potential of machine learning (ML), deep learning (DL), and artificial intelligence (AI) in precision medicine. The goal of this work is to show progress in ML in digital health, to exemplify future needs and trends, and to identify any essential prerequisites of AI and ML for precision health.

RECENT FINDINGS

High-throughput technologies are delivering growing volumes of biomedical data, such as large-scale genome-wide sequencing assays; libraries of medical images; or drug perturbation screens of healthy, developing, and diseased tissue. Multi-omics data in biomedicine is deep and complex, offering an opportunity for data-driven insights and automated disease classification. Learning from these data will open our understanding and definition of healthy baselines and disease signatures. State-of-the-art applications of deep neural networks include digital image recognition, single-cell clustering, and virtual drug screens, demonstrating breadths and power of ML in biomedicine.

SUMMARY

Significantly, AI and systems biology have embraced big data challenges and may enable novel biotechnology-derived therapies to facilitate the implementation of precision medicine approaches.

MetSigDis: a manually curated resource for the metabolic signatures of diseases

Complex diseases cannot be understood only on the basis of single gene, single mRNA transcript or single protein but the effect of their collaborations. The combination consequence in molecular level can be captured by the alterations of metabolites. With the rapidly developing of biomedical instruments and analytical platforms, a large number of metabolite signatures of complex diseases were identified and documented in the literature. Biologists' hardship in the face of this large amount of papers recorded metabolic signatures of experiments' results calls for an automated data repository. Therefore, we developed MetSigDis aiming to provide a comprehensive resource of metabolite alterations in various diseases. MetSigDis is freely available at http://www.bio-annotation.cn/MetSigDis/. By reviewing hundreds of publications, we collected 6849 curated relationships between 2420 metabolites and 129 diseases across eight species involving Homo sapiens and model organisms. All of these relationships were used in constructing a metabolite disease network (MDN). This network displayed scale-free characteristics according to the degree distribution (power-law distribution with R2 = 0.909), and the subnetwork of MDN for interesting diseases and their related metabolites can be visualized in the Web. The common alterations of metabolites reflect the metabolic similarity of diseases, which is measured using Jaccard index. We observed that metabolite-based similar diseases are inclined to share semantic associations of Disease Ontology. A human disease network was then built, where a node represents a disease, and an edge indicates similarity of pair-wise diseases. The network validated the observation that linked diseases based on metabolites should have more overlapped genes.

Assessing key decisions for transcriptomic data integration in biochemical networks

To gain insights into complex biological processes, genome-scale data (e.g., RNA-Seq) are often overlaid on biochemical networks. However, many networks do not have a one-to-one relationship between genes and network edges, due to the existence of isozymes and protein complexes. Therefore, decisions must be made on how to overlay data onto networks. For example, for metabolic networks, these decisions include (1) how to integrate gene expression levels using gene-protein-reaction rules, (2) the approach used for selection of thresholds on expression data to consider the associated gene as "active", and (3) the order in which these steps are imposed. However, the influence of these decisions has not been systematically tested. We compared 20 decision combinations using a transcriptomic dataset across 32 tissues and showed that definition of which reaction may be considered as active (i.e., reactions of the genome-scale metabolic network with a non-zero expression level after overlaying the data) is mainly influenced by thresholding approach used. To determine the most appropriate decisions, we evaluated how these decisions impact the acquisition of tissue-specific active reaction lists that recapitulate organ-system tissue groups. These results will provide guidelines to improve data analyses with biochemical networks and facilitate the construction of context-specific metabolic models.

Human Systems Biology and Metabolic Modelling: A Review-From Disease Metabolism to Precision Medicine

In cell and molecular biology, metabolism is the only system that can be fully simulated at genome scale. Metabolic systems biology offers powerful abstraction tools to simulate all known metabolic reactions in a cell, therefore providing a snapshot that is close to its observable phenotype. In this review, we cover the 15 years of human metabolic modelling. We show that, although the past five years have not experienced large improvements in the size of the gene and metabolite sets in human metabolic models, their accuracy is rapidly increasing. We also describe how condition-, tissue-, and patient-specific metabolic models shed light on cell-specific changes occurring in the metabolic network, therefore predicting biomarkers of disease metabolism. We finally discuss current challenges and future promising directions for this research field, including machine/deep learning and precision medicine. In the omics era, profiling patients and biological processes from a multiomic point of view is becoming more common and less expensive. Starting from multiomic data collected from patients and N-of-1 trials where individual patients constitute different case studies, methods for model-building and data integration are being used to generate patient-specific models. Coupled with state-of-the-art machine learning methods, this will allow characterizing each patient's disease phenotype and delivering precision medicine solutions, therefore leading to preventative medicine, reduced treatment, and in silico clinical trials.

Predicting gastrointestinal drug effects using contextualized metabolic models

Gastrointestinal side effects are among the most common classes of adverse reactions associated with orally absorbed drugs. These effects decrease patient compliance with the treatment and induce undesirable physiological effects. The prediction of drug action on the gut wall based on in vitro data solely can improve the safety of marketed drugs and first-in-human trials of new chemical entities. We used publicly available data of drug-induced gene expression changes to build drug-specific small intestine epithelial cell metabolic models. The combination of measured in vitro gene expression and in silico predicted metabolic rates in the gut wall was used as features for a multilabel support vector machine to predict the occurrence of side effects. We showed that combining local gut wall-specific metabolism with gene expression performs better than gene expression alone, which indicates the role of small intestine metabolism in the development of adverse reactions. Furthermore, we reclassified FDA-labeled drugs with respect to their genetic and metabolic profiles to show hidden similarities between seemingly different drugs. The linkage of xenobiotics to their transcriptomic and metabolic profiles could take pharmacology far beyond the usual indication-based classifications.

HUME: large-scale detection of causal genetic factors of adverse drug reactions

Motivation

Adverse drug reactions are one of the major factors that affect the wellbeing of patients and financial costs of healthcare systems. Genetic variations of patients have been shown to be a key factor in the occurrence and severity of many ADRs. However, the large number of confounding drugs and genetic biomarkers for each adverse reaction case demands a method that evaluates all potential genetic causes of ADRs simultaneously.

Results

To address this challenge, we propose HUME, a multi-phase algorithm that recommends genetic factors for ADRs that are causally supported by the patient record data. HUME consists of the construction of a network from co-prevalence between significant genetic biomarkers and ADRs, a link score phase for predicting candidate relations based on the Adamic-Adar measure, and a causal refinement phase based on multiple hypothesis testing of quasi experimental designs for evaluating evidence and counter evidence of candidate relations in the patient records.

Supplementary information

Supplementary data are available at Bioinformatics online.

Integrated human-virus metabolic stoichiometric modelling predicts host-based antiviral targets against Chikungunya, Dengue and Zika viruses

Current and reoccurring viral epidemic outbreaks such as those caused by the Zika virus illustrate the need for rapid development of antivirals. Such development would be facilitated by computational approaches that can provide experimentally testable predictions for possible antiviral strategies. To this end, we focus here on the fact that viruses are directly dependent on their host metabolism for reproduction. We develop a stoichiometric, genome-scale metabolic model that integrates human macrophage cell metabolism with the biochemical demands arising from virus production and use it to determine the virus impact on host metabolism and vice versa. While this approach applies to any host–virus pair, we first apply it to currently epidemic viruses Chikungunya, Dengue and Zika in this study. We find that each of these viruses causes specific alterations in the host metabolic flux towards fulfilling their biochemical demands as predicted by their genome and capsid structure. Subsequent analysis of this integrated model allows us to predict a set of host reactions, which, when constrained, inhibit virus production. We show that this prediction recovers known targets of existing antiviral drugs, specifically those targeting nucleotide production, while highlighting a set of hitherto unexplored reactions involving both amino acid and nucleotide metabolic pathways, with either broad or virus-specific antiviral potential. Thus, this computational approach allows rapid generation of experimentally testable hypotheses for novel antiviral targets within a host.

Recon3D enables a three-dimensional view of gene variation in human metabolism

Genome-scale network reconstructions have helped uncover the molecular basis of metabolism. Here we present Recon3D, a computational resource that includes three-dimensional (3D) metabolite and protein structure data and enables integrated analyses of metabolic functions in humans. We use Recon3D to functionally characterize mutations associated with disease, and identify metabolic response signatures that are caused by exposure to certain drugs. Recon3D represents the most comprehensive human metabolic network model to date, accounting for 3,288 open reading frames (representing 17% of functionally annotated human genes), 13,543 metabolic reactions involving 4,140 unique metabolites, and 12,890 protein structures. These data provide a unique resource for investigating molecular mechanisms of human metabolism. Recon3D is available at http://vmh.life.

Predicting Dynamic Metabolic Demands in the Photosynthetic Eukaryote Chlorella vulgaris

Phototrophic organisms exhibit a highly dynamic proteome, adapting their biomass composition in response to diurnal light/dark cycles and nutrient availability. Here, we used experimentally determined biomass compositions over the course of growth to determine and constrain the biomass objective function (BOF) in a genome-scale metabolic model of Chlorella vulgaris UTEX 395 over time. Changes in the BOF, which encompasses all metabolites necessary to produce biomass, influence the state of the metabolic network thus directly affecting predictions. Simulations using dynamic BOFs predicted distinct proteome demands during heterotrophic or photoautotrophic growth. Model-driven analysis of extracellular nitrogen concentrations and predicted nitrogen uptake rates revealed an intracellular nitrogen pool, which contains 38% of the total nitrogen provided in the medium for photoautotrophic and 13% for heterotrophic growth. Agreement between flux and gene expression trends was determined by statistical comparison. Accordance between predicted flux trends and gene expression trends was found for 65% of multisubunit enzymes and 75% of allosteric reactions. Reactions with the highest agreement between simulations and experimental data were associated with energy metabolism, terpenoid biosynthesis, fatty acids, nucleotides, and amino acid metabolism. Furthermore, predicted flux distributions at each time point were compared with gene expression data to gain new insights into intracellular compartmentalization, specifically for transporters. A total of 103 genes related to internal transport reactions were identified and added to the updated model of C. vulgaris, iCZ946, thus increasing our knowledgebase by 10% for this model green alga.

Elucidation of complexity and prediction of interactions in microbial communities

Microorganisms engage in complex interactions with other members of the microbial community, higher organisms as well as their environment. However, determining the exact nature of these interactions can be challenging due to the large number of members in these communities and the manifold of interactions they can engage in. Various omic data, such as 16S rRNA gene sequencing, shotgun metagenomics, metatranscriptomics, metaproteomics and metabolomics, have been deployed to unravel the community structure, interactions and resulting community dynamics in situ. Interpretation of these multi-omic data often requires advanced computational methods. Modelling approaches are powerful tools to integrate, contextualize and interpret experimental data, thus shedding light on the underlying processes shaping the microbiome. Here, we review current methods and approaches, both experimental and computational, to elucidate interactions in microbial communities and to predict their responses to perturbations.

Metabolic profiling of triple-negative breast cancer cells reveals metabolic vulnerabilities

Background

Among breast cancers, the triple-negative breast cancer (TNBC) subtype has the worst prognosis with no approved targeted therapies and only standard chemotherapy as the backbone of systemic therapy. Unique metabolic changes in cancer progression provide innovative therapeutic opportunities. The receptor tyrosine kinases (RTKs) epidermal growth factor receptor (EGFR), and MET receptor are highly expressed in TNBC, making both promising therapeutic targets. RTK signaling profoundly alters cellular metabolism by increasing glucose consumption and subsequently diverting glucose carbon sources into metabolic pathways necessary to support the tumorigenesis. Therefore, detailed metabolic profiles of TNBC subtypes and their response to tyrosine kinase inhibitors may identify therapeutic sensitivities.

Methods

We quantified the metabolic profiles of TNBC cell lines representing multiple TNBC subtypes using gas chromatography mass spectrometry. In addition, we subjected MDA-MB-231, MDA-MB-468, Hs578T, and HCC70 cell lines to metabolic flux analysis of basal and maximal glycolytic and mitochondrial oxidative rates. Metabolic pool size and flux measurements were performed in the presence and absence of the MET inhibitor, INC280/capmatinib, and the EGFR inhibitor, erlotinib. Further, the sensitivities of these cells to modulators of core metabolic pathways were determined. In addition, we annotated a rate-limiting metabolic enzymes library and performed a siRNA screen in combination with MET or EGFR inhibitors to validate synergistic effects.

Results

TNBC cell line models displayed significant metabolic heterogeneity with respect to basal and maximal metabolic rates and responses to RTK and metabolic pathway inhibitors. Comprehensive systems biology analysis of metabolic perturbations, combined siRNA and tyrosine kinase inhibitor screens identified a core set of TCA cycle and fatty acid pathways whose perturbation sensitizes TNBC cells to small molecule targeting of receptor tyrosine kinases.

Conclusions

Similar to the genomic heterogeneity observed in TNBC, our results reveal metabolic heterogeneity among TNBC subtypes and demonstrate that understanding metabolic profiles and drug responses may prove valuable in targeting TNBC subtypes and identifying therapeutic susceptibilities in TNBC patients. Perturbation of metabolic pathways sensitizes TNBC to inhibition of receptor tyrosine kinases. Such metabolic vulnerabilities offer promise for effective therapeutic targeting for TNBC patients.

Electronic supplementary material

The online version of this article (doi:10.1186/s40170-017-0168-x) contains supplementary material, which is available to authorized users.

Analysis of free text in electronic health records for identification of cancer patient trajectories

With an aging patient population and increasing complexity in patient disease trajectories, physicians are often met with complex patient histories from which clinical decisions must be made. Due to the increasing rate of adverse events and hospitals facing financial penalties for readmission, there has never been a greater need to enforce evidence-led medical decision-making using available health care data. In the present work, we studied a cohort of 7,741 patients, of whom 4,080 were diagnosed with cancer, surgically treated at a University Hospital in the years 2004-2012. We have developed a methodology that allows disease trajectories of the cancer patients to be estimated from free text in electronic health records (EHRs). By using these disease trajectories, we predict 80% of patient events ahead in time. By control of confounders from 8326 quantified events, we identified 557 events that constitute high subsequent risks (risk > 20%), including six events for cancer and seven events for metastasis. We believe that the presented methodology and findings could be used to improve clinical decision support and personalize trajectories, thereby decreasing adverse events and optimizing cancer treatment.

Proteomic analysis of red blood cells and the potential for the clinic: what have we learned so far?

INTRODUCTION

Red blood cells (RBC) are the most abundant host cells in the human body. Mature erythrocytes are devoid of nuclei and organelles and have always been regarded as circulating 'bags of hemoglobin'. The advent of proteomics has challenged this assumption, revealing unanticipated complexity and novel roles for RBCs not just in gas transport, but also in systemic metabolic homeostasis in health and disease. Areas covered: In this review we will summarize the main advancements in the field of discovery mode and redox/quantitative proteomics with respect to RBC biology. We thus focus on translational/clinical applications, such as transfusion medicine, hematology (e.g. hemoglobinopathies) and personalized medicine. Synergy of omics technologies - especially proteomics and metabolomics - are highlighted as a hallmark of clinical metabolomics applications for the foreseeable future. Expert commentary: The introduction of advanced proteomics technologies, especially quantitative and redox proteomics, and the integration of proteomics data with omics information gathered through orthogonal technologies (especially metabolomics) promise to revolutionize many biomedical areas, from hematology and transfusion medicine to personalized medicine and clinical biochemistry.

Reconciled rat and human metabolic networks for comparative toxicogenomics and biomarker predictions

The laboratory rat has been used as a surrogate to study human biology for more than a century. Here we present the first genome-scale network reconstruction of Rattus norvegicus metabolism, iRno, and a significantly improved reconstruction of human metabolism, iHsa. These curated models comprehensively capture metabolic features known to distinguish rats from humans including vitamin C and bile acid synthesis pathways. After reconciling network differences between iRno and iHsa, we integrate toxicogenomics data from rat and human hepatocytes, to generate biomarker predictions in response to 76 drugs. We validate comparative predictions for xanthine derivatives with new experimental data and literature-based evidence delineating metabolite biomarkers unique to humans. Our results provide mechanistic insights into species-specific metabolism and facilitate the selection of biomarkers consistent with rat and human biology. These models can serve as powerful computational platforms for contextualizing experimental data and making functional predictions for clinical and basic science applications.

The rat is a widely-used model for human biology, but we must be aware of metabolic differences. Here, the authors reconstruct the genome-scale metabolic network of the rat, and after reconciling it with an improved human metabolic model, demonstrate the power of the models to integrate toxicogenomics data, providing species-specific biomarker predictions in response to a panel of drugs.

A Multi-scale Computational Platform to Mechanistically Assess the Effect of Genetic Variation on Drug Responses in Human Erythrocyte Metabolism

Progress in systems medicine brings promise to addressing patient heterogeneity and individualized therapies. Recently, genome-scale models of metabolism have been shown to provide insight into the mechanistic link between drug therapies and systems-level off-target effects while being expanded to explicitly include the three-dimensional structure of proteins. The integration of these molecular-level details, such as the physical, structural, and dynamical properties of proteins, notably expands the computational description of biochemical network-level properties and the possibility of understanding and predicting whole cell phenotypes. In this study, we present a multi-scale modeling framework that describes biological processes which range in scale from atomistic details to an entire metabolic network. Using this approach, we can understand how genetic variation, which impacts the structure and reactivity of a protein, influences both native and drug-induced metabolic states. As a proof-of-concept, we study three enzymes (catechol-O-methyltransferase, glucose-6-phosphate dehydrogenase, and glyceraldehyde-3-phosphate dehydrogenase) and their respective genetic variants which have clinically relevant associations. Using all-atom molecular dynamic simulations enables the sampling of long timescale conformational dynamics of the proteins (and their mutant variants) in complex with their respective native metabolites or drug molecules. We find that changes in a protein's structure due to a mutation influences protein binding affinity to metabolites and/or drug molecules, and inflicts large-scale changes in metabolism.

Personalized Kinetic Models for Predictive Healthcare

Metabolomics data are used to parameterize individual-specific kinetic models of metabolism to predict medically relevant parameters underpinning disease states or outcomes.

The SIDER database of drugs and side effects

Unwanted side effects of drugs are a burden on patients and a severe impediment in the development of new drugs. At the same time, adverse drug reactions (ADRs) recorded during clinical trials are an important source of human phenotypic data. It is therefore essential to combine data on drugs, targets and side effects into a more complete picture of the therapeutic mechanism of actions of drugs and the ways in which they cause adverse reactions. To this end, we have created the SIDER (‘Side Effect Resource’, http://sideeffects.embl.de) database of drugs and ADRs. The current release, SIDER 4, contains data on 1430 drugs, 5880 ADRs and 140 064 drug–ADR pairs, which is an increase of 40% compared to the previous version. For more fine-grained analyses, we extracted the frequency with which side effects occur from the package inserts. This information is available for 39% of drug–ADR pairs, 19% of which can be compared to the frequency under placebo treatment. SIDER furthermore contains a data set of drug indications, extracted from the package inserts using Natural Language Processing. These drug indications are used to reduce the rate of false positives by identifying medical terms that do not correspond to ADRs.