Detection Of Drug-drug Interactions Research Articles

Significant increases in detection capability for assessment of organic anion transporting polypeptide-mediated drug-drug interaction in cynomolgus monkeys by modifying pretreatment of Coproporphyrin I and III, and glycochenodeoxycholate-3-sulfate in plasma

BackgroundCoproporphyrin I (CP-I), Coproporphyrin III (CP-III), and glycochenodeoxycholate-3-sulfate (GCDCA-S) act as endogenous substrates of Organic Anion Transporting Polypeptide (OATP) 1B and have been considered for application in OATP1B-mediated drug‒drug interaction (DDI) risk assessments. Prior assays of the endogenous OATP substrates might exhibit reduced DDI detection capability and possibly overlook low DDI risk. We pioneered a simultaneous assay of the three substrates in monkey plasma using ultra-performance liquid chromatography with tandem mass spectrometry (UPLC-MS/MS) and applied it to monkey studies to identify lower DDI risk. ResultsThe methodology development indicated that precursors of CP-I/III were oxidized to form CP-I/III, diminishing the detection capability in DDI risk assessments. A precursor-eliminated analytical (PEA) method was developed to eliminate the precursors through solid-phase extraction. This method aimed to prevent the oxidation of CP-I/III precursors by incorporating edaravone. For comparison, a precursor-oxidized analytical (POA) method was also developed, wherein the precursors of CP-I/III were fully oxidized to CP-I/III. The PEA method achieved high sensitivity for CP-I/III and GCDCA-S, with lower quantification limits of 0.01 ng mL-1 and 0.5 ng mL-1, respectively. Both methods ensured that the validation parameters met the acceptance criteria. The two methods were applied to a monkey study, with CP-I/III showcasing notably enhanced DDI detection capabilities through the novel PEA method in comparison to the POA method. SignificanceThis study’s methodology has future implications for OATP-mediated DDI risk assessment using endogenous substrates. The novel PEA method can identify lower OATP-mediated DDI risks for drugs that the current methods cannot detect. Our method is likely applicable in clinical settings, and its utility should be assessed in clinical trials.

Prescription Precision: A Comprehensive Review of Intelligent Prescription Systems.

Intelligent Prescription Systems (IPS) represent a promising frontier in healthcare, offering the potential to optimize medication selection, dosing, and monitoring tailored to individual patient needs. This comprehensive review explores the current landscape of IPS, encompassing various technological approaches, applications, benefits, and challenges. IPS leverages advanced computational algorithms, machine learning techniques, and big data analytics to analyze patient-specific factors, such as medical history, genetic makeup, biomarkers, and lifestyle variables. By integrating this information with evidence-based guidelines, clinical decision support systems, and real-time patient data, IPS generates personalized treatment recommendations that enhance therapeutic outcomes while minimizing adverse effects and drug interactions. Key components of IPS include predictive modeling, Drug-Drug Interaction detection, adverse event prediction, dose optimization, and medication adherence monitoring. These systems offer clinicians invaluable decision-support tools to navigate the complexities of medication management, particularly in the context of polypharmacy and chronic disease management. While IPS holds immense promise for improving patient care and reducing healthcare costs, several challenges must be addressed. These include data privacy and security concerns, interoperability issues, integration with existing electronic health record systems, and clinician adoption barriers. Additionally, the regulatory landscape surrounding IPS requires clarification to ensure compliance with evolving healthcare regulations. Despite these challenges, the rapid advancements in artificial intelligence, data analytics, and digital health technologies are driving the continued evolution and adoption of IPS. As precision medicine gains momentum, IPS is poised to play a central role in revolutionizing medication management, ultimately leading to more effective, personalized, and patient-centric healthcare delivery.

Drug interaction detection and glycemic control – telepharmacy’s role in diabetes management: before-after study

Introduction: Diabetes management faces intertwined challenges like polypharmacy and adherence complexities, hindering effective care delivery. Telepharmacy emerges as a promising solution, revolutionizing diabetes care by enhancing pharmacists’ pivotal role. However, a tailored model for Indonesia is still lacking. Purpose: This study aimed to assess the potential effectiveness of a telepharmacy application in improving the detection of drug-drug interactions and glycemic control. Methods: We conducted an observational before-and-after study among type 2 diabetic patients in Indonesia. Results: We noted a significant surge in detected drug interactions. Patients counseled through telepharmacy had 3.653 times higher likelihood of achieving good glycemic control. Conclusion: The study underscores the positive impact of telepharmacy on drug interaction detection and glycemic control among Indonesian diabetics.

Detection and management of clinically relevant drug-drug interactions with direct oral anticoagulants: an intervention study in community pharmacies

IntroductionDirect oral anticoagulants (DOACs) are increasingly used and can be involved in clinically relevant drug-drug interactions (DDIs) that increase the risk of major bleeding or thromboembolism. Skilled drug interaction management is essential to ensure safe and effective use of DOACs. In this study, we aimed to investigate the impact of the detection and management of DDIs with DOACs in a real-life community pharmacy setting on the pharmacotherapy of DOAC users. MethodsWe conducted an intervention study in 201 community pharmacies in Belgium. On random days, patients purchasing DOACs or drugs known to interact with them were screened. When a DDI with the DOAC was detected, the pharmacist contacted the prescribing physician to discuss the management of the interaction. A previously developed practice-oriented DDI list accompanied by management plans for ambulatory care was used for both screening and management of the DDIs. ResultsIn total, 751 patients were included, among whom 875 DDIs were identified, primarily pharmacodynamic DDIs (95.7 %). Predominant interacting drug classes included selective serotonin or serotonin and norepinephrine reuptake inhibitors (32.9 %), antiplatelets (30.9 %), and non-steroidal anti-inflammatory drugs (28.9 %). In 43.0 % of DDIs, an intervention was decided upon. At three-month follow-up, proposed pharmacotherapy changes had been implemented in 79.1 % of these DDIs. ConclusionsThis study demonstrates that active screening and management of DDIs with DOACs in community pharmacies, in close collaboration with prescribing physicians, resulted in changes in pharmacotherapy in a substantial number of patients. This may contribute significantly to the safer utilisation of DOACs in high-risk populations.

GGI-DDI: Identification for key molecular substructures by granule learning to interpret predicted drug–drug interactions

Deep learning-based approaches have achieved promising performance in predicting drug–drug interactions (DDIs). Nevertheless, a significant drawback of these approaches is their limited interpretability, hindering their practical applicability for clinicians. Based on current research findings, drug interactions frequently arise from specific substructures or functional groups present in drugs. To enhance the interpretability of deep learning models, we propose a novel end-to-end learning approach that employs granular computing to identify pivotal substructures instead of using conventional atom-based or predefined molecular fingerprint methods to predict DDIs. We refer to this model as ”GGI-DDI” (Granule–Granule Interaction for Drug–Drug Interaction). In this method, drugs are granulated into a set of coarser granules that represent the key substructures or functional groups of drugs. Subsequently, the detection of DDIs occurs through the examination of interactions among these granules, aligning more closely with human cognitive patterns. We conducted thorough experiments on the TWOSIDES dataset, and the results show that GGI-DDI achieved impeccable accuracy in predicting DDIs. We compared GGI-DDI to state-of-the-art baseline models including DDIMDL, GoGNN, DNN, STNN-DDI and GMPNN-CS, GGI-DDI almost consistently outperforms the baselines across all metrics in terms of Accuracy (Acc), Area under the receiver operating characteristic (Auc), Area under precision recall curve (Aupr) and Precision (Pre) in both transductive and inductive scenarios. Finally, we provide case studies to illustrate how GGI-DDI can effectively reveal important substructure pairs across drugs about a specific DDI type, offering insights into the underlying mechanism of these interactions. We find that this interpretability can serve as valuable guidance in the advancement of novel drug development and poly-drug therapy strategies.

A metabolomic analysis of sensitivity and specificity of 23 previously proposed biomarkers for renal transporter-mediated drug-drug interactions.

Endogenous biomarkers are discussed as tools for detection of drug-drug interactions mediated by renal transport proteins, such as organic cation transporter 2 (OCT2), multidrug and toxin extrusion proteins (MATE1 and MATE2-K) and organic anion transporters (OAT1 and OAT3). Whereas sensitivity of some endogenous biomarkers against at least one clinical transporter inhibitor has frequently been shown, intrastudy comparisons of the extent of effects of inhibitors on different biomarkers are frequently lacking. Moreover, in vivo specificity of such discussed biomarkers has frequently not been studied. We therefore investigated changes of ten previously described putative biomarkers for inhibition of OCT2/MATEs, as well as 15 previously described putative biomarkers for OATs in human plasma and urine samples of healthy volunteers in response to treatment with four inhibitors of transport proteins [verapamil (P-glycoprotein), rifampin (organic anion transporting polypeptides), cimetidine (OCT2/MATEs), and probenecid (OATs)]. Two of the putative biomarkers had been suggested for both OCT2/MATEs and OATs. All substances were unequivocally identified in an untargeted metabolomics assay. The OCT2/MATE biomarkers choline and trimethylamine N-oxide were both sensitive and specific (median log2-fold changes -1.18 in estimated renal elimination and -0.85 in urinary excretion, respectively). For renal OATs, indoleacetyl glutamine and indoleacetic acid (median log2-fold changes -3.77 and -2.85 in estimated renal elimination, respectively) were the candidates for sensitive and specific biomarkers with the most extensive change, followed by taurine, indolelactic acid and hypoxanthine. This comprehensive study adds further knowledge on sensitivity and specificity of 23 previously described biomarkers of renal OCT2/MATE- and OAT-mediated drug-drug interactions.

EMSI-BERT: Asymmetrical Entity-Mask Strategy and Symbol-Insert Structure for Drug–Drug Interaction Extraction Based on BERT

Drug-drug interaction (DDI) extraction has seen growing usage of deep models, but their effectiveness has been restrained by limited domain-labeled data, a weak representation of co-occurring entities, and poor adaptation of downstream tasks. This paper proposes a novel EMSI-BERT method for drug–drug interaction extraction based on an asymmetrical Entity-Mask strategy and a Symbol-Insert structure. Firstly, the EMSI-BERT method utilizes the asymmetrical Entity-Mask strategy to address the weak representation of co-occurring entity information using the drug entity dictionary in the pre-training BERT task. Secondly, the EMSI-BERT method incorporates four symbols to distinguish different entity combinations of the same input sequence and utilizes the Symbol-Insert structure to address the week adaptation of downstream tasks in the fine-tuning stage of DDI classification. The experimental results showed that EMSI-BERT for DDI extraction achieved a 0.82 F1-score on DDI-Extraction 2013, and it improved the performances of the multi-classification task of DDI extraction and the two-classification task of DDI detection. Compared with baseline Basic-BERT, the proposed pre-training BERT with the asymmetrical Entity-Mask strategy could obtain better effects in downstream tasks and effectively limit “Other” samples’ effects. The model visualization results illustrated that EMSI-BERT could extract semantic information at different levels and granularities in a continuous space.

A SYSTEMATIC REVIEW OF SOFTWARE AND RESOURCES AVAILABLE IN THE PHARMACEUTICAL SECTOR

The whole pharmaceutical area (pharmacy industry, pharmacy billing, clinical trials, pharmacovigilance, and drug-drug interaction detection) needed various innovative and scientific solutions to solve the problems. A systematic review was performed on various software applications widely used in the pharmaceutical industry, including clinical trials, pharmacy billing, pharmacovigilance, the detection of drug interactions that assist the healthcare professional. Our review concluded that computer software has created a significant role in the pharmaceutical industry, pharmacy billing, and herbal industry. The computer software and programmes aid in reducing manufacturing costs, manufacturing errors, medical errors, and clinical errors.The software performs various safety functions, supplies on-screen access to pertinent drug and patient information, manages data, and assists in the product formulation process. The programmed safeguards can be modified to meet changing requirements.

Prediction of drug-drug interaction events using graph neural networks based feature extraction

The prevalence of multi_drug therapies has been increasing in recent years, particularly among the elderly who are suffering from several diseases. However, unexpected Drug_Drug interaction (DDI) can cause adverse reactions or critical toxicity, which puts patients in danger. As the need for multi_drug treatment increases, it's becoming increasingly necessary to discover DDIs. Nevertheless, DDIs detection in an extensive number of drug pairs, both in-vitro and in-vivo, is costly and laborious. Therefore, DDI identification is one of the most concerns in drug-related researches. In this paper, we propose GNN-DDI, a deep learning-based method for predicting DDI-associated events in two stages. In the first stage, we collect the drugs information from different sources and then integrate them through the formation of an attributed heterogeneous network and generate a drug embedding vector based on different drug interaction types and drug attributes. In the second stage, we aggregate the representation vectors then predictions of the DDIs and their events are performed through a deep multi-model framework. Various evaluation results show that the proposed method can outperform state-of-the methods in the prediction of drug-drug interaction-associated events. The experimental results indicate that producing the drug's representations based on different drug interaction types and attributes is efficient and effective and can better show the intrinsic characteristics of a drug.

Prediction of Drug-Drug Interaction Using an Attention-Based Graph Neural Network on Drug Molecular Graphs.

The treatment of complex diseases by using multiple drugs has become popular. However, drug-drug interactions (DDI) may give rise to the risk of unanticipated adverse effects and even unknown toxicity. Therefore, for polypharmacy safety it is crucial to identify DDIs and explore their underlying mechanisms. The detection of DDI in the wet lab is expensive and time-consuming, due to the need for experimental research over a large volume of drug combinations. Although many computational methods have been developed to predict DDIs, most of these are incapable of predicting potential DDIs between drugs within the DDI network and new drugs from outside the DDI network. In addition, they are not designed to explore the underlying mechanisms of DDIs and lack interpretative capacity. Thus, here we propose a novel method of GNN-DDI to predict potential DDIs by constructing a five-layer graph attention network to identify k-hops low-dimensional feature representations for each drug from its chemical molecular graph, concatenating all identified features of each drug pair, and inputting them into a MLP predictor to obtain the final DDI prediction score. The experimental results demonstrate that our GNN-DDI is suitable for each of two DDI predicting scenarios, namely the potential DDIs among known drugs in the DDI network and those between drugs within the DDI network and new drugs from outside DDI network. The case study indicates that our method can explore the specific drug substructures that lead to the potential DDIs, which helps to improve interpretability and discover the underlying interaction mechanisms of drug pairs.

Facilitating primary care non-antiretroviral drug prescribing in people living with HIV: The 'THINK ARV' initiative.

Older people living with HIV (PLWH) have higher rates of multimorbidity, polypharmacy and an associated increased risk of potential drug-drug interactions (DDIs). We describe the development, implementation and evaluation of an intervention to increase community prescribers' access to specialist prescribing advice. Phase One: a survey evaluating General Practitioners' (GPs') knowledge of, and confidence detecting DDIs affecting PLWH, was circulated to eight General Practices in one UK city. Phase Two: co-production was used to develop the THINK ARV intervention for prescribers in city-wide General Practices: a dedicated mobile phone and e-mail advice service staffed by HIV specialist pharmacists. Queries were audited for 6months pre- and post-intervention. A user-satisfaction survey was emailed to enquirers. Phase One: 42 GPs responded, of whom 62% requested further support identifying DDIs among PLWH. Phase Two: the number of queries received increased from 25 (6months before 'THINK ARV' launch) to 63 in the following 6months (152% increase). 94% of the queries were specifically about DDIs. Increasing community prescribers' access to specialist telephone and e-mail advice resulted in increased awareness and detection of DDIs. Similar interventions could be embedded within different healthcare settings to optimise medicines and avoid potential patient harm.

Is It Time to Use Modeling of Cellular Transporter Homeostasis to Inform Drug-Drug Interaction Studies: Theoretical Considerations.

Mathematical modeling has been an important tool in pharmaceutical research for 50 + years and there is increased emphasis over the last decade on using modeling to improve the efficiency and effectiveness of drug development. In an earlier commentary, we applied a multiscale model linking 6 scales (whole body, tumor, vasculature, cell, spatial location, time), together with literature data on nanoparticle and tumor properties, to demonstrate the effects of nanoparticle particles on systemic disposition. The current commentary used a 4-scale model (cell membrane, intracellular organelles, spatial location, time) together with literature data on the intracellular processing of membrane receptors and transporters to demonstrate disruption of transporter homeostasis can lead to drug-drug interaction (DDI) between victim drug (VD) and perpetrator drug (PD), including changes in the area-under-concentration-time-curve of VD in cells that are considered significant by the US Food and Drug Administration (FDA). The model comprised 3 computational components: (a) intracellular transporter homeostasis, (b) pharmacokinetics of extracellular and intracellular VD/PD concentrations, and (c) pharmacodynamics of PD-induced stimulation or inhibition of an intracellular kinetic process. Model-based simulations showed that (a) among the five major endocytic processes, perturbation of transporter internalization or recycling led to the highest incidence and most extensive DDI, with minor DDI for perturbing transporter synthesis and early-to-late endosome and no DDI for perturbing transporter degradation and (b) three experimental conditions (spatial transporter distribution in cells, VD/PD co-incubation time, extracellular PD concentrations) were determinants of DDI detection. We propose modeling is a useful tool for hypothesis generation and for designing experiments to identify potential DDI; its application further aligns with the model-informed drug development paradigm advocated by FDA.

A Comprehensive Review of Computational Methods For Drug-Drug Interaction Detection.

The detection of drug-drug interactions (DDIs) is a crucial task for drug safety surveillance, which provides effective and safe co-prescriptions of multiple drugs. Since laboratory researches are often complicated, costly and time-consuming, it's urgent to develop computational approaches to detect drug-drug interactions. In this paper, we conduct a comprehensive review of state-of-the-art computational methods falling into three categories: literature-based extraction methods, machine learning-based prediction methods and pharmacovigilance-based data mining methods. Literature-based extraction methods detect DDIs from published literature using natural language processing techniques; machine learning-based prediction methods build prediction models based on the known DDIs in databases and predict novel ones; pharmacovigilance-based data mining methods usually apply statistical techniques on various electronic data to detect drug-drug interaction signals. We first present the taxonomy of drug-drug interaction detection methods and provide the outlines of three categories of methods. Afterwards, we respectively introduce research backgrounds and data sources of three categories, and illustrate their representative approaches as well as evaluation metrics. Finally, we discuss the current challenges of existing methods and highlight potential opportunities for future directions.

SumGNN: multi-typed drug interaction prediction via efficient knowledge graph summarization.

Thanks to the increasing availability of drug-drug interactions (DDI) datasets and large biomedical knowledge graphs (KGs), accurate detection of adverse DDI using machine learning models becomes possible. However, it remains largely an open problem how to effectively utilize large and noisy biomedical KG for DDI detection. Due to its sheer size and amount of noise in KGs, it is often less beneficial to directly integrate KGs with other smaller but higher quality data (e.g. experimental data). Most of existing approaches ignore KGs altogether. Some tries to directly integrate KGs with other data via graph neural networks with limited success. Furthermore most previous works focus on binary DDI prediction whereas the multi-typed DDI pharmacological effect prediction is more meaningful but harder task. To fill the gaps, we propose a new method SumGNN: knowledge summarization graph neural network, which is enabled by a subgraph extraction module that can efficiently anchor on relevant subgraphs from a KG, a self-attention based subgraph summarization scheme to generate reasoning path within the subgraph, and a multi-channel knowledge and data integration module that utilizes massive external biomedical knowledge for significantly improved multi-typed DDI predictions. SumGNN outperforms the best baseline by up to 5.54%, and performance gain is particularly significant in low data relation types. In addition, SumGNN provides interpretable prediction via the generated reasoning paths for each prediction. The code is available in Supplementary Material. Supplementary data are available at Bioinformatics online.

Potential drug-drug interactions of antiretrovirals and antimicrobials detected by three databases

Standard treatment for HIV infection involves a combination of antiretrovirals. Additionally, opportunistic infections in HIV infected patients require further antimicrobial medications that might cause drug-drug interactions (DDIs). The objective of this study was to to compare the recognition of DDIs between antiretrovirals and antimicrobials by three proprietary databases and evaluate their concordance. 114 items of antiretrovirals and antimicrobials from the National List of Essential Medicines of Thailand 2018 were used in the study. However, 21 items were not recognised by Micromedex, Drugs.com, and Liverpool HIV interactions. Only 93 items were available for the detection of potential DDIs by the three databases. Potential DDIs detected from the three databases included 292 pairs. Liverpool showed the highest number of DDIs with 285 pairs compared with 259 pairs by drugs.com and 133 pairs by Micromedex. Regarding the severity classifications, Liverpool reported 10% Contraindicated; Micromedex reported 14% contraindicated and 59% major; Drugs.com reported 21% major. The Fleiss’ kappa agreements were fair to poor among the three databases, higher agreement was observed for DDIs classified as severe. This study highlights the need to harmonize the evaluation and interpretation of DDI risk in order to produce standardized information to support prescribers.

Does DDI-Predictor Help Pharmacists to Detect Drug-Drug Interactions and Resolve Medication Issues More Effectively?

The characterization of drug-drug interactions (DDIs) may require the use of several different tools, such as the thesaurus issued by our national health agency (i.e., ANSM), the metabolic pathways table from the Geneva University Hospital (GUH), and DDI-Predictor (DDI-P). We sought to (i) compare the three tools’ respective abilities to detect DDIs in routine clinical practice and (ii) measure the pharmacist intervention rate (PIR) and physician acceptance rate (PAR) associated with the use of DDI-P. The three tools’ respective DDI detection rates (in %) were measured. The PIRs and PARs were compared by using the area under the curve ratio given by DDI-P (RAUC) and applying a chi-squared test. The DDI detection rates differed significantly: 40.0%, 76.5%, and 85.2% for ANSM (The National Agency for the Safety of Medicines and Health Products), GUH and DDI-P, respectively (p < 0.0001). The PIR differed significantly according to the DDI-P’s RAUC: 90.0%, 44.2% and 75.0% for RAUC ≤ 0.5; RAUC 0.5–2 and RAUC > 2, respectively (p < 0.001). The overall PAR was 85.1% and did not appear to depend on the RAUC category (p = 0.729). Our results showed that more pharmacist interventions were issued when details of the strength of the DDI were available. The three tools can be used in a complementary manner, with a view to refining medication adjustments.

A span-graph neural model for overlapping entity relation extraction in biomedical texts.

Entity relation extraction is one of the fundamental tasks in biomedical text mining, which is usually solved by the models from natural language processing. Compared with traditional pipeline methods, joint methods can avoid the error propagation from entity to relation, giving better performances. However, the existing joint models are built upon sequential scheme, and fail to detect overlapping entity and relation, which are ubiquitous in biomedical texts. The main reason is that sequential models have relatively weaker power in capturing long-range dependencies, which results in lower performance in encoding longer sentences. In this article, we propose a novel span-graph neural model for jointly extracting overlapping entity relation in biomedical texts. Our model treats the task as relation triplets prediction, and builds the entity-graph by enumerating possible candidate entity spans. The proposed model captures the relationship between the correlated entities via a span scorer and a relation scorer, respectively, and finally outputs all valid relational triplets. Experimental results on two biomedical entity relation extraction tasks, including drug-drug interaction detection and protein-protein interaction detection, show that the proposed method outperforms previous models by a substantial margin, demonstrating the effectiveness of span-graph-based method for overlapping relation extraction in biomedical texts. Further in-depth analysis proves that our model is more effective in capturing the long-range dependencies for relation extraction compared with the sequential models. Related codes are made publicly available at http://github.com/Baxelyne/SpanBioER.

DPDDI: a deep predictor for drug-drug interactions

BackgroundThe treatment of complex diseases by taking multiple drugs becomes increasingly popular. However, drug-drug interactions (DDIs) may give rise to the risk of unanticipated adverse effects and even unknown toxicity. DDI detection in the wet lab is expensive and time-consuming. Thus, it is highly desired to develop the computational methods for predicting DDIs. Generally, most of the existing computational methods predict DDIs by extracting the chemical and biological features of drugs from diverse drug-related properties, however some drug properties are costly to obtain and not available in many cases.ResultsIn this work, we presented a novel method (namely DPDDI) to predict DDIs by extracting the network structure features of drugs from DDI network with graph convolution network (GCN), and the deep neural network (DNN) model as a predictor. GCN learns the low-dimensional feature representations of drugs by capturing the topological relationship of drugs in DDI network. DNN predictor concatenates the latent feature vectors of any two drugs as the feature vector of the corresponding drug pairs to train a DNN for predicting the potential drug-drug interactions. Experiment results show that, the newly proposed DPDDI method outperforms four other state-of-the-art methods; the GCN-derived latent features include more DDI information than other features derived from chemical, biological or anatomical properties of drugs; and the concatenation feature aggregation operator is better than two other feature aggregation operators (i.e., inner product and summation). The results in case studies confirm that DPDDI achieves reasonable performance in predicting new DDIs.ConclusionWe proposed an effective and robust method DPDDI to predict the potential DDIs by utilizing the DDI network information without considering the drug properties (i.e., drug chemical and biological properties). The method should also be useful in other DDI-related scenarios, such as the detection of unexpected side effects, and the guidance of drug combination.

Comparison of Signal Detection Algorithms Based on Frequency Statistical Model for Drug-Drug Interaction Using Spontaneous Reporting Systems.

Adverse events (AEs) caused by polypharmacy have recently become a clinical problem, and it is important to monitor the safety profile of drug-drug interactions (DDIs). Mining signals using the spontaneous reporting systems is a very effective method for single drug-induced AE monitoring as well as early detection of DDIs. The objective of this study was to compare signal detection algorithms for DDIs based on frequency statistical models. Five frequency statistical models: the Ω shrinkage measure, additive (risk difference), multiplicative (risk ratio), combination risk ratio, and chi-square statistics models were compared using the Japanese Adverse Drug Event Report (JADER) database as the spontaneous reporting system in Japan. The drugs targeted for the survey are all registered and classified as "suspect drugs" in JADER, and the AEs targeted for this study were the same as those in a previous study on Stevens-Johnson syndrome (SJS). Of 3924 pairs that reported SJS, the number of signals detected by the Ω shrinkage measure, additive, multiplicative, combination risk ratio, and chi-square statistics models was 712, 3298, 2252, 739, and 1289 pairs, respectively. Among the five models, the Ω shrinkage measure model showed the most conservative signal detection tendency. Specifically, caution should be exercised when the number of reports is low because results differ depending on the statistical models. This study will contribute to the selection of appropriate statistical models to detect signals of potential DDIs.

Structure-Based Drug-Drug Interaction Detection via Expressive Graph Convolutional Networks and Deep Sets (Student Abstract)

In this work, we proposed a DDI detection method based on molecular structures using graph convolutional networks and deep sets. We proposed a more discriminative convolutional layer compared to conventional GCN and achieved permutation invariant prediction without losing the capability of capturing complicated interactions.