Network-based prediction of drug combinations

Invariant set theory for predicting potential failure of antibiotic cycling

Collateral sensitivity, where resistance to one drug confers heightened sensitivity to another, offers a promising strategy for combating antimicrobial resistance, yet predicting resultant evolutionary dynamics remains a significant challenge. We propose here a mathematical model that integrates fitness trade-offs and adaptive landscapes to predict the evolution of collateral sensitivity pathways, providing insights into optimizing sequential drug therapies. Our approach embeds collateral information into a network of switched systems, allowing us to abstract the effects of sequential antibiotic exposure on antimicrobial resistance. We analyze the system stability at disease-free equilibrium and employ set-control theory to tailor therapeutic windows. Consequently, we propose a computational algorithm to identify effective sequential therapies to counter antibiotic resistance. By leveraging our theory with data on collateral sensivity interactions, we predict scenarios that may prevent bacterial escape for chronic Pseudomonas aeruginosa infections.

Predicting the risk of ischemic stroke in patients with atrial fibrillation using heterogeneous drug-protein-disease network-based deep learning

Current risk assessment models for predicting ischemic stroke (IS) in patients with atrial fibrillation (AF) often fail to account for the effects of medications and the complex interactions between drugs, proteins, and diseases. We developed an interpretable deep learning model, the AF-Biological-IS-Path (ABioSPath), to predict one-year IS risk in AF patients by integrating drug-protein-disease pathways with real-world clinical data. Using a heterogeneous multilayer network, ABioSPath identifies mechanisms of drug actions and the propagation of comorbid diseases. By combining mechanistic pathways with patient-specific characteristics, the model provides individualized IS risk assessments and identifies potential molecular pathways involved. We utilized the electronic health record data from 7859 AF patients, collected between January 2008 and December 2009 across 43 hospitals in Hong Kong. ABioSPath outperformed baseline models in all evaluation metrics, achieving an AUROC of 0.7815 (95% CI: 0.7346-0.8283), a positive predictive value of 0.430, a negative predictive value of 0.870, a sensitivity of 0.500, a specificity of 0.885, an average precision of 0.409, and a Brier score of 0.195. Cohort-level analysis identified key proteins, such as CRP, REN, and PTGS2, within the most common pathways. Individual-level analysis further highlighted the importance of PIK3/Akt and cytokine and chemokine signaling pathways and identified IS risks associated with less-studied drugs like prochlorperazine maleate. ABioSPath offers a robust, data-driven approach for IS risk prediction, requiring only routinely collected clinical data without the need for costly biomarkers. Beyond IS, the model has potential applications in screening risks for other diseases, enhancing patient care, and providing insights for drug development.

DeepCCDS: Interpretable Deep Learning Framework for Predicting Cancer Cell Drug Sensitivity through Characterizing Cancer Driver Signals

Accurate characterization of cellular states is the foundation for precise prediction of drug sensitivity in cancer cell lines, which in turn is fundamental to realizing precision oncology. However, current deep learning approaches have limitations in characterizing cellular states. They rely solely on isolated genetic markers, overlooking the complex regulatory networks and cellular mechanisms that underlie drug responses. To address this limitation, this work proposes DeepCCDS, a Deep learning framework for Cancer Cell Drug Sensitivity prediction through Characterizing Cancer Driver Signals. DeepCCDS incorporates a prior knowledge network to characterize cancer driver signals, building upon the self-supervised neural network framework. The signals can reflect key mechanisms influencing cancer cell development and drug response, enhancing the model's predictive performance and interpretability. DeepCCDS has demonstrated superior performance in predicting drug sensitivity compared to previous state-of-the-art approaches across multiple datasets. Benefiting from integrating prior knowledge, DeepCCDS exhibits powerful feature representation capabilities and interpretability. Based on these feature representations, we have identified embedding features that could potentially be used for drug screening in new indications. Further, this work demonstrates the applicability of DeepCCDS on solid tumor samples from The Cancer Genome Atlas. This work believes integrating DeepCCDS into clinical decision-making processes can potentially improve the selection of personalized treatment strategies for cancer patients.

Network-based multi-class classifier to identify optimized gene networks for acute leukemia cell line classification

Unraveling the genetic regulatory networks that underlie diseases is essential for comprehending the intricate mechanisms of these conditions. While various computational strategies were developed, the approaches in the existing studies concerning network-based prediction and classification are based on the pre-estimated gene networks. However, the gene network that is pre-estimated fails to yield biologically meaningful explanations for classifying cell lines into particular clinical states. The reason for this limitation is the lack of inclusion of any information about the clinical status of cell lines during the process of network estimation. To achieve effective cell line classification and ensure the biological validity of the cell lines classification, we develop a computational strategy referred to as GRN-multiClassifier for network-based multi-class classification. The GRN-multiClassifier estimates gene network in a manner that simultaneously minimizes both the network estimation error and the negative log-likelihood function of multinomial logistic regression. That is, our strategy estimates optimized gene network to enable the multi-class classification of cell lines into specific clinical conditions. Monte Carlo simulations demonstrate the efficacy of the GRN-multiClassifier. We applied our strategy to network-based classification of acute leukemia cell lines into three distinct categories of acute leukemia. Our strategy shows outstanding performance in the classification of acute leukemia cell lines. The results for the acute leukemia marker identification are strongly supported by existing literature. The implications of our findings suggest that potential pathways involving the inhibition of ACTB and the molecular interactions between "HBA1&HBB," "HBB&HBA1," "IGKV1-5&IGHV4-31," "IGHV4-31&IGKV1-5," "HLA-DRA&CD74" and "ACTB&ACTB" could offer significant insights into the underlying mechanism of acute leukemia.

Understanding dosage effects of traditional Chinese medicine using network analysis

Background

Traditional Chinese Medicine (TCM) prescriptions are complex, multi-botanical drug systems in which dosage critically influences therapeutic efficacy. While network pharmacology is widely used to analyze TCM mechanisms, existing methods ignore the dosage of botanical drugs, a key limitation that may skew predictions. This study investigates how integrating dosage data alters network analysis outputs, addressing a fundamental gap in understanding TCM's dosage-dependent effects.

Methods

Our analysis compared dosage-weighted and traditional non-dosage network approaches across 94 traditional Chinese medicine (TCM) prescriptions. We developed four custom indicators to quantify differences throughout the network pipeline: Dedis (input distance difference), DeSD (input standard deviation difference), DeDT (drug target prediction difference), and DePy (pathway prediction difference). The interrelationships among these indicators were examined to indicate when dosage adjustments influence predictions. A detailed case study further demonstrated the impact of dosage modifications on predictive outcomes.

Results

Among the indicators with inputs difference, Dedis, but not DeSD, exhibited a statistically significant relationship with output predictions, with target differences (DeDT) ranging from 0% to 68.9% and pathway differences (DePy) ranging from 0% to 74.6%. The interrelationships between these indicators were visualized using a clock model representation. The case study further demonstrated the impact of dosage on network outputs, revealing dosage refined both the predicted drug targets for individual botanical drugs and the subsequent pathway analysis results.

Conclusion

Our study demonstrated that dosage significantly influences the outcomes of network analysis, with Dedis serving as a reliable indicator of whether such changes would occur. Specifically, changes resulting from dosage-dependent refinement of both drug target prediction and pathway analysis were observed.

Protective effect of deinoxanthin in sorafenib-induced nephrotoxicity in rats with the hepatocellular carcinoma model

Sorafenib is a synthetic compound and an orally administered multichines inhibitor that targets growth signaling and angiogenesis. It is widely recognized as the standard of care for advanced hepatocellular carcinoma (HCC) but has toxic side effects. Deinoxanthin, purified from the radioresistant bacterium Deinococcus radiodurans, has strong antioxidant characteristics. In this study, the protective effect of deinoxanthin against sorafenib-induced nephrotoxicity was investigated in a rat model of hepatocellular carcinoma. In this regard, the expressions of DDAH1, KIM1, and INOS genes were examined, histopathological and immunohistochemical analyses were performed, and various parameters such as SOD, MDA, GST, CAT, TAS, and TOS were tested biochemically. BUN and creatinine levels were measured in renal tissues. RT-qPCR, Western blot, and ELISA methods were used for all these analyses. As a result, the analyses show that deinoxanthin, which has a high antioxidant capacity, reduces kidney injury and can be used as a protective agent. The primary objective of this study is to evaluate the potential of deinoxanthin as a protective agent against the nephrotoxic side effects of sorafenib in HCC. Our study identified the potential synergistic effects of sorafenib and deinoxanthin on nephrotoxicity in rats with hepatocellular carcinoma.

Decoding herbal combination models through systematic strategies: insights from target information and traditional Chinese medicine clinical theory

Traditional Chinese medicine (TCM) utilizes intricate herbal formulations that exemplify the principles of compatibility and synergy. However, the rapid proliferation of herbal data has resulted in redundant information, complicating the understanding of their potential mechanisms. To address this issue, we first established a comprehensive database that encompasses 992 herbs, 18 681 molecules, and 2168 targets. Consequently, we implemented a multi-network strategy based on a core information screening method to elucidate the highly intertwined relationships among the targets of various herbs and to refine herbal target information. Within a non-redundant network framework, separation and overlap analysis demonstrated that the networking of herbs preserves essential clinical information, including their properties, meridians, and therapeutic classifications. Furthermore, two notable trends emerged from the statistical analyses of classical TCM formulas: the separation of herbs and the overlap between herbs and diseases. This phenomenon is termed the herbal combination model (HCM), validated through statistical analyses of two representative case studies: the common cold and rheumatoid arthritis. Additionally, in vivo and in vitro experiments with the new formula YanChuanQin (YanHuSuo-Corydalis Rhizoma, ChuanWu-Aconiti Radix, and QinJiao-Gentianae Macrophyllae Radix) for acute gouty arthritis further support the HCM. Overall, this computational method provides a systematic network strategy for exploring herbal combinations in complex and poorly understood diseases from a non-redundant perspective.

Predicting rare drug-drug interaction events with dual-granular structure-adaptive and pair variational representation

Adverse drug-drug interaction events (DDIEs) pose serious risks to patient safety, yet rare but severe interactions remain challenging to identify due to limited clinical data. Existing computational methods rely heavily on abundant samples, failing to identify rare DDIEs. Here we introduce RareDDIE, a metric-based meta-learning model that employs a dual-granular structure-driven pair variational representation to enhance rare DDIE prediction. To further address the challenge of zero-shot DDIE identification, we develop the Biological Semantic Transferring (BST) module, integrating large-scale sentence embeddings to form the ZetaDDIE variant. Our model outperforms existing methods in few-sample and zero-sample settings. Furthermore, we verify that knowledge transfer from DDIE can improve drug synergy predictions, surpassing existing models. Case studies on antiplatelet activity reduction and non-small cell lung cancer drug synergy further illustrate the practical value of RareDDIE. By analyzing the meta-knowledge construction process, we provide interpretability into the model's decision-making. This work establishes an effective computational framework for rare DDIE prediction, leveraging meta-learning and knowledge transfer to overcome key challenges in data-limited scenarios.

AI-driven discovery of synergistic drug combinations against pancreatic cancer

Pancreatic cancer treatment often relies on multi-drug regimens, but optimal combinations remain elusive. This study evaluates predictive approaches to identify synergistic drug combinations using a dataset from the National Center for Advancing Translational Sciences (NCATS). Screening 496 combinations of 32 anticancer compounds against the PANC-1 cells experimentally determined the degree of synergism and antagonism. Three research groups (NCATS, University of North Carolina, and Massachusetts Institute of Technology) leverage these data to apply machine learning (ML) approaches, predicting synergy across 1.6 million combinations. Of the 88 tested, 51 show synergy, with graph convolutional networks achieving the best hit rate and random forest the highest precision. Beyond highlighting the potential of ML, this work delivers 307 experimentally validated synergistic combinations, demonstrating its practical impact in treating pancreatic cancer.

A novel model of glioblastoma recurrence to identify therapeutic vulnerabilities

Glioblastoma remains incurable and recurs in all patients. Here we design and characterize a novel induced-recurrence model in which mice xenografted with primary patient-derived glioma initiating/stem cells (GIC) are treated with a therapeutic regimen closely recapitulating patient standard of care, followed by monitoring until tumours recur (induced recurrence patient-derived xenografts, IR-PDX). By tracking in vivo tumour growth, we confirm the patient specificity and initial efficacy of treatment prior to recurrence. Availability of longitudinally matched pairs of primary and recurrent GIC enabled patient-specific evaluation of the fidelity with which the model recapitulated phenotypes associated with the true recurrence. Through comprehensive multi-omic analyses, we show that the IR-PDX model recapitulates aspects of genomic, epigenetic, and transcriptional state heterogeneity upon recurrence in a patient-specific manner. The accuracy of the IR-PDX enabled both novel biological insights, including the positive association between glioblastoma recurrence and levels of ciliated neural stem cell-like tumour cells, and the identification of druggable patient-specific therapeutic vulnerabilities. This proof-of-concept study opens the possibility for prospective precision medicine approaches to identify target-drug candidates for treatment at glioblastoma recurrence.

Identification and Drug Screening of Single Cells from Human Tumors on Semiconductor Chip for Cancer Precision Medicine

Drug screening of primary tumor cells directly assesses the drug efficacy on specific tumors, promoting personalized cancer treatment. The application of a microfluidic platform has realized drug screening using a limited amount of biopsy samples for cancer precision medicine. However, all the techniques face an inevitable issue of not all the primary tumor cells being cancer cells. Here, a system is introduced that integrates single-cell identification and drug screening on one semiconductor chip so that both drug efficacy on cancer cells and drug toxicity on noncancerous cells can be obtained simultaneously. An integrated circuit is built on the semiconductor chip for single-cell electric impedance sensing (IC-ECIS) of ultra-weak signals for distinguishing cancer cells from noncancerous cells without affecting cell vitality. Single-cell identification is validated using breast, lung, and liver cell lines as well as liver cancer specimens from clinical patients. The accuracy on commercial cell lines is ≈80%, and the diagnostic results of tumor tissues are consistent with clinical pathology results. Drug screening is run on the same chip after single cell identification for dual evaluation of drug efficacy and toxicity in both breast cancer models and clinical liver cancer patients. The on-chip drug screening is confirmed with off-chip counterpart experiments in breast cell lines. The effectiveness or ineffectiveness of a drug screened on the IC-ECIS chip demonstrated consistency in the presence or absence of specific mutations in the drug-related genes determined via exome sequencing of individual liver tumors, validating the method for precision medicine.

Tumors and their microenvironments: Learning from pediatric brain pathologies

Early clues to tumors and their microenvironments come from embryonic development. Here we review the literature and consider whether the embryonic brain and its pathologies can serve as a better model. Among embryonic organs, the brain is the most heterogenous and complex, with multiple lineages leading to wide spectrum of cell states and types. Its dysregulation promotes neurodevelopmental brain pathologies and pediatric tumors. Embryonic brain pathologies point to the crucial importance of spatial heterogeneity over time, akin to the tumor microenvironment. Tumors dedifferentiate through genetic mutations and epigenetic modulations; embryonic brains differentiate through epigenetic modulations. Our innovative review proposes learning developmental brain pathologies to target tumor evolution-and vice versa. We describe ways through which tumor pharmacology can learn from embryonic brains and their pathologies, and how learning tumor, and its microenvironment, can benefit targeting neurodevelopmental pathologies. Examples include pediatric low-grade versus high-grade brain tumors as in rhabdomyosarcomas and gliomas.

Uncovering the anti-cancer mechanism of cucurbitacin D against colorectal cancer through network pharmacology and molecular docking

Colorectal cancer is a significant global health challenge due to chemoresistance, necessitating new treatments. Cucurbitacin D, with its anti-cancer properties, shows promise, but its effects on colorectal cancer are not well understood. We investigated the impact of cucurbitacin D on colorectal cancer cell lines using MTT assays and Annexin V/7-AAD staining followed by flow cytometry for apoptosis analysis. Public databases helped identify cucurbitacin D and colorectal cancer-related gene targets for network pharmacology analysis. Protein-protein interaction networks were constructed using STRING and analyzed in Cytoscape. Gene ontology and KEGG pathway enrichment analyses were performed using ClueGo. Molecular docking studies were conducted via Autodock Vina and visualized in Discovery Studio. Western blot assessed protein expression changes in key targets under cucurbitacin D. Cucurbitacin D dose-dependently reduced colorectal cancer cell viability and induced apoptosis. Network pharmacology pinpointed crucial targets like STAT3, AKT1, CCND1, and CASP3. Molecular docking confirmed strong interactions with these targets. Enrichment analysis highlighted involvement in the 'PI3K-AKT,' 'JAK-STAT,' and 'ErbB' signaling pathways. These findings suggest cucurbitacin D as a potential anti-colorectal cancer agent, demonstrating significant effects on cell viability and apoptosis, and engaging critical cancer-related pathways, making it a promising candidate for further colorectal cancer therapeutic research.

CPDP: Contrastive Protein-Drug Pre-Training for Novel Drug Discovery

Novel drug discovery and repositioning remain critical challenges in biomedical research, requiring accurate prediction of drug-target interactions (DTIs). We propose the CPDP framework, which builds upon existing biomedical representation models and integrates contrastive learning with multi-dimensional representations of proteins and drugs to predict DTIs. By aligning the representation space, CPDP enables GNN-based methods to achieve zero-shot learning capabilities, allowing for accurate predictions of unseen drug data. This approach enhances DTI prediction performance, particularly for novel drugs not included in the BioHNs dataset. Experimental results demonstrate CPDP's high accuracy and strong generalization ability in predicting novel biological entities while maintaining effectiveness for traditional drug repositioning tasks.

TCnet: A Novel Strategy to Predict Target Combination of Alzheimer's Disease via Network-Based Methods

Alzheimer's disease (AD) is a complex neurodegenerative disorder with an unclear pathogenesis; the traditional ″single gene-single target-single drug″ strategy is insufficient for effective treatment. This study explores a novel strategy for the multitarget therapy of AD by integrating multiomics data and employing network analysis. Different from conventional single-target methods, TCnet adopts a mechanism-driven strategy, utilizing multiomics data to decompose disease mechanisms, construct potential target combinations, and prioritize the optimal combinations using a scoring function. TCnet not only advances our understanding of disease mechanisms but also facilitates large-scale drug screening. This approach was further employed to screen active compounds from Huang-Lian-Jie-Du-Tang (HLJDT), identifying quercetin as a candidate targeting GSK3β and ADAM17. Subsequent in vitro experiments confirmed the neuroprotective and anti-inflammatory effects of quercetin. Overall, TCnet offers a promising approach for predicting target combinations and provides new insights and directions for drug discovery in AD.

Modular-Based Synergetic Mechanisms of Jasminoidin and Ursodeoxycholic Acid in Cerebral Ischemia Therapy

Objectives: Jasminoidin (JA) and ursodeoxycholic acid (UA) have been shown to exert synergistic effects on cerebral ischemia (CI) therapy, but the underlying mechanisms remain to be elucidated. Objective: To elucidate the synergistic mechanisms involved in the combined use of JA and UA (JU) for CI therapy using a driver-induced modular screening (DiMS) strategy. Methods: Network proximity and topology-based approaches were used to identify synergistic modules and driver genes from an anti-ischemic microarray dataset (ArrayExpress, E-TABM-662). A middle cerebral artery occlusion/reperfusion (MCAO/R) model was established in 30 Sprague Dawley rats, divided into sham, vehicle, JA (25 mg/mL), UA (7 mg/mL), and JU (JA:UA = 1:1) groups. After 90 minutes of ischemia, infarct volume and neurological deficit scores were evaluated. Western blotting was performed 24 h after administration to validate key protein changes. Results: Six, eleven, and four drug-responsive On_modules were identified for JA, UA, and JU, respectively. Three synergistic modules (Sy-modules, JU-Mod-7, 8, and 10) and 12 driver genes (e.g., NRF1, FN1, CUL3) were identified, mainly involving the PI3K-Akt and MAPK pathways and regulation of the actin cytoskeleton. JA and UA synergistically reduced infarct volume and neurological deficit score (2.5, p < 0.05) in MCAO/R rats. In vivo studies demonstrated that JU suppressed the expression of CUL3, FN1, and ITGA4, while it increased that of NRF1. Conclusions: JU acts synergistically on CI-reperfusion injury by regulating FN1, CUL3, ITGA4, and NRF1 and inducing the PI3K-Akt, MAPK, and actin cytoskeleton pathways. DiMS provides a new approach to uncover mechanisms of combination therapies.

Caution when using network partners for target identification in drug discovery

Identifying novel, high-yield drug targets is challenging and often results in a high failure rate. However, recent data indicate that leveraging human genetic evidence to identify and validate these targets significantly increases the likelihood of success in drug development. Two recent papers from Open Targets claimed that around half of US Food and Drug Administration-approved drugs had targets with direct human genetic evidence. By expanding target identification to include protein network partners-molecules in physical contact-the proportion of drug targets with genetic evidence support increased to two-thirds. However, the efficacy of using these network partners for target identification was not formally tested. To address this, we tested the approach on a list of robust positive control genes. We used the IntAct database to find physically interacting proteins of genes identified by exome-wide association studies (ExWASs), genome-wide association studies (GWASs) combined with a locus-to-gene mapping algorithm called the Effector Index, and Genetic Priority Score (GPS), which integrated eight genetic features with drug indications from the Open Targets and SIDER databases. We assessed how accurately including interacting genes with the ExWAS-, Effector Index-, and GPS-selected genes identified positive controls, focusing on precision, sensitivity, and specificity. Our results indicated that although molecular interactions led to higher sensitivity in identifying positive control genes, their practical application is limited by low precision. Expanding genetically identified targets to include network partners using IntAct did not increase the likelihood of identifying drug targets across the 412 tested traits, suggesting that such results should be interpreted with caution.

Utilization of precision medicine digital twins for drug discovery in Alzheimer's disease

Alzheimer's disease (AD) presents significant challenges in drug discovery and development due to its complex and poorly understood pathology and etiology. Digital twins (DTs) are recently developed virtual real-time representations of physical entities that enable rapid assessment of the bidirectional interaction between the virtual and physical domains. With recent advances in artificial intelligence (AI) and the growing accumulation of multi-omics and clinical data, application of DTs in healthcare is gaining traction. Digital twin technology, in the form of multiscale virtual models of patients or organ systems, can track health status in real time with continuous feedback, thereby driving model updates that enhance clinical decision-making. Here, we posit an additional role for DTs in drug discovery, with particular utility for complex diseases like AD. In this review, we discuss salient challenges in AD drug development, including complex disease pathology and comorbidities, difficulty in early diagnosis, and the current high failure rate of clinical trials. We also review DTs and discuss potential applications for predicting AD progression, discovering biomarkers, identifying new drug targets and opportunities for drug repurposing, facilitating clinical trials, and advancing precision medicine. Despite significant hurdles in this area, such as integration and standardization of dynamic medical data and issues of data security and privacy, DTs represent a promising approach for revolutionizing drug discovery in AD.

UVB-induced HaCat cell damage and Myricaria Paniculata's molecular effects

The Qinghai‒Tibet Plateau, the "Roof of the World" in China, has high altitude, low pressure, thin air, little rain, long sunshine, and snow cover, causing 80-90% more Ultraviolet (UV) reflectance and greater skin UV exposure at high altitudes. Myricaria paniculata, a Tibetan plant growing at 2000-4500 m, has anti-inflammatory, antioxidant, and immune-boosting effects and can protect skin cells from Ultraviolet B (UVB)damage. The protective effects of Myricaria paniculata compounds against UVB-induced HaCat cell damage were explored. Samples were divided into normal, model, and treatment groups (seven compounds). First, the cell viability and apoptosis rates of each group were measured, along with the levels of factors such as Reactive oxygen species (ROS) and Superoxide dismutase (SOD). Network pharmacology analysis and molecular docking were subsequently performed. This study revealed that the compound enhanced cell survival, inhibited apoptosis, reduced ROS and Malondialdehyde (MDA) levels, and increased SOD activity. It also lowered the levels of Interleukin-6 (IL-6), Tumor Necrosis Factor-α (TNF-α), and Aspartate protein hydrolase 3 containing cysteine (Caspase-3). An analysis of the intersection between the 218 targets of the seven compounds found in Myricaria paniculata and the 1002 targets associated with skin inflammation revealed 59 common targets, with key targets including TNF and others. GO and KEGG analyses suggested the involvement of metabolic pathways. Seven core targets related to skin inflammation in Myricaria paniculata were identified by molecular docking. In addition, its compounds rhamnetin, rhamnocitrin, ferulic acid and kaempferol have good binding activity with TNF, PTGS2, EGFR and MMP9 targets. The Tibetan medicine Myricaria paniculata had a certain protective effect on UVB-induced HaCat cell damage.

A new paradigm for drug discovery in the treatment of complex diseases: drug discovery and optimization

In the past, the drug research and development has predominantly followed a "single target, single disease" model. However, clinical data show that single-target drugs are difficult to interfere with the complete disease network, are prone to develop drug resistance and low safety in clinical use. The proposal of multi-target drug therapy (also known as "cocktail therapy") provides a new approach for drug discovery, which can affect the disease and reduce adverse reactions by regulating multiple targets. Natural products are an important source for multi-target innovative drug development, and more than half of approved small molecule drugs are related to natural products. However, there are many challenges in the development process of natural products, such as active drug screening, target identification and preclinical dosage optimization. Therefore, how to develop multi-target drugs with good drug resistance from natural products has always been a challenge. This article summarizes the applications and shortcomings of related technologies such as natural product bioactivity screening, clarify the mode of action of the drug (direct/indirect target), and preclinical dose optimization. Moreover, in response to the challenges faced by natural products in the development process and the trend of interdisciplinary and multi-technology integration, and a multi-target drug development strategy of "active substances - drug action mode - drug optimization" is proposed to solve the key challenges in the development of natural products from multiple dimensions and levels.

Computational drug repurposing: approaches, evaluation of in silico resources and case studies

Repurposing of existing drugs for new indications has attracted substantial attention owing to its potential to accelerate drug development and reduce costs. Hundreds of computational resources such as databases and predictive platforms have been developed that can be applied for drug repurposing, making it challenging to select the right resource for a specific drug repurposing project. With the aim of helping to address this challenge, here we overview computational approaches to drug repurposing based on a comprehensive survey of available in silico resources using a purpose-built drug repurposing ontology that classifies the resources into hierarchical categories and provides application-specific information. We also present an expert evaluation of selected resources and three drug repurposing case studies implemented within the Horizon Europe REMEDi4ALL project to demonstrate the practical use of the resources. This comprehensive Review with expert evaluations and case studies provides guidelines and recommendations on the best use of various in silico resources for drug repurposing and establishes a basis for a sustainable and extendable drug repurposing web catalogue.

The Importance of Murine Models in Determining In Vivo Pharmacokinetics, Safety, and Efficacy in Antimalarial Drug Discovery

New chemical entities are constantly being investigated towards antimalarial drug discovery, and they require animal models for toxicity and efficacy testing. Murine models show physiological similarities to humans and are therefore indispensable in the search for novel antimalarial drugs. They provide a preclinical basis (following in vitro assessments of newly identified lead compounds) for further assessment in the drug development pipeline. Specific mouse strains, non-humanized and humanized, have successfully been infected with rodent Plasmodium species and the human Plasmodium species, respectively. Infected mice provide a platform for the assessment of treatment options being sought. In vivo pharmacokinetic evaluations are necessary when determining the fate of potential antimalarials in addition to the efficacy assessment of these chemical entities. This review describes the role of murine models in the drug development pipeline. It also explains some in vivo pharmacokinetic, safety, and efficacy parameters necessary for making appropriate choices of lead compounds in antimalarial drug discovery. Despite the advantages of murine models in antimalarial drug discovery, certain limitations are also highlighted.

A fast approach for structural and evolutionary analysis based on energetic profile protein comparison

In structural bioinformatics, the efficiency of predicting protein similarity, function, and evolutionary relationships is crucial. Our approach proposed herein leverages protein energy profiles derived from a knowledge-based potential, deviating from traditional methods relying on structural alignment or atomic distances. This method assigns unique energy profiles to individual proteins, facilitating rapid comparative analysis for both structural similarities and evolutionary relationships across various hierarchical levels. Our study demonstrates that energy profiles contain substantial information about protein structure at class, fold, superfamily, and family levels. Notably, these profiles accurately distinguish proteins across species, illustrated by the classification of coronavirus spike glycoproteins and bacteriocin proteins. Introducing a separation measure based on energy profile similarity, our method shows significant correlation with a network-based approach, emphasizing the potential of energy profiles as efficient predictors for drug combinations with faster computational requirements. Our key insight is that the sequence-based energy profile strongly correlates with structure-derived energy, enabling rapid and efficient protein comparisons based solely on sequences.

Integrating transcriptomic data with a novel drug efficacy prediction model for TCM active compound discovery

Identifying the active natural compounds remains a challenge for drug discovery, and new algorithms need to be developed to predict active ingredients from complex natural products. Here, we proposed Meta-DEP, a Meta-paths-based Drug Efficacy Prediction based on drug-protein-disease heterogeneity network, where Meta-paths contain all the shortest paths between drug targets and disease-related proteins in the network and drug efficacy is measured by a predictive score according to drug disease network proximity. Experiments show that Meta-DEP performs better than traditional network topology analysis on drug-disease interaction prediction task. Further investigations demonstrate that the key targets identified by Meta-DEP for drug efficacy are consistent with clinical pharmacological evidence. To prove that Meta-DEP can be used to discover active natural compounds, we apply it to predict the relationship between the monomeric components of traditional Chinese medicine included in the TCMSP database and diseases. Results indicate that Meta-DEP can accurately predict most of the drug-disease pairs included in the TCMSP database. In addition, biological experiments are directly used to demonstrate that Meta-DEP can mined active compound from traditional Chinese medicine with integrating disease transcriptomic data. Overall, the model developed in this study provides new impetus for driving the natural compound into innovative lead molecule. Code and data are available at https://github.com/t9lex/Meta-DEP .

PathSynergy: a deep learning model for predicting drug synergy in liver cancer

Cancer is a major public health problem while liver cancer is the main cause of global cancer-related deaths. The previous study demonstrates that the 5-year survival rate for advanced liver cancer is only 30%. Few of the first-line targeted drugs including sorafenib and lenvatinib are available, which often develop resistance. Drug combination therapy is crucial for improving the efficacy of cancer therapy and overcoming resistance. However, traditional methods for discovering drug synergy are costly and time consuming. In this study, we developed a novel predicting model PathSynergy by integrating drug feature data, cell line data, drug-target interactions, and signaling pathways. PathSynergy combined the advantages of graph neural networks and pathway map mapping. Comparing with other baseline models, PathSynergy showed better performance in model classification, accuracy, and precision. Excitingly, six Food and Drug Administration (FDA)-approved drugs including pimecrolimus, topiramate, nandrolone_decanoate, fluticasone propionate, zanubrutinib, and levonorgestrel were predicted and validated to show synergistic effects with sorafenib or lenvatinib against liver cancer for the first time. In general, the PathSynergy model provides a new perspective to discover synergistic combinations of drugs and has broad application potential in the fields of drug discovery and personalized medicine.

SensitiveCancerGPT: Leveraging Generative Large Language Model on Structured Omics Data to Optimize Drug Sensitivity Prediction

Objective

The fast accumulation of vast pharmacogenomics data of cancer cell lines provide unprecedented opportunities for drug sensitivity prediction (DSP), a crucial prerequisite for the advancement of precision oncology. Recently, Generative Large Language Models (LLM) have demonstrated performance and generalization prowess across diverse tasks in the field of natural language processing (NLP). However, the structured format of the pharmacogenomics data poses challenge for the utility of LLM in DSP. Therefore, the objective of this study is multi-fold: to adapt prompt engineering for structured pharmacogenomics data toward optimizing LLM's DSP performance, to evaluate LLM's generalization in real-world DSP scenarios, and to compare LLM's DSP performance against that of state-of-the-science baselines.

Methods

We systematically investigated the capability of the Generative Pre-trained Transformer (GPT) as a DSP model on four publicly available benchmark pharmacogenomics datasets, which are stratified by five cancer tissue types of cell lines and encompass both oncology and non-oncology drugs. Essentially, the predictive landscape of GPT is assessed for effectiveness on the DSP task via four learning paradigms: zero-shot learning, few-shot learning, fine-tuning and clustering pretrained embeddings. To facilitate GPT in seamlessly processing the structured pharmacogenomics data, domain-specific novel prompt engineering is employed by implementing three prompt templates (i.e., Instruction, Instruction-Prefix, Cloze) and integrating pharmacogenomics-related features into the prompt. We validated GPT's performance in diverse real-world DSP scenarios: cross-tissue generalization, blind tests, and analyses of drug-pathway associations and top sensitive/resistant cell lines. Furthermore, we conducted a comparative evaluation of GPT against multiple Transformer-based pretrained models and existing DSP baselines.

Results

Extensive experiments on the pharmacogenomics datasets across the five tissue cohorts demonstrate that fine-tuning GPT yields the best DSP performance (28% F1 increase, p-value= 0.0003) followed by clustering pretrained GPT embeddings (26% F1 increase, p-value= 0.0005), outperforming GPT in-context learning (i.e., few-shot). However, GPT in the zero-shot setting had a big F1 gap, resulting in the worst performance. Within the scope of prompt engineering, performance enhancement was achieved by directly instructing GPT about the DSP task and resorting to a concise context format (i.e., instruction-prefix), leading to F1 performance gain of 22% (p-value=0.02); while incorporation of drug-cell line prompt context derived from genomics and/or molecular features further boosted F1 score by 2%. Compared to state-of-the-science DSP baselines, GPT significantly asserted superior mean F1 performance (16% gain, p-value<0.05) on the GDSC dataset. In the cross-tissue analysis, GPT showcased comparable generalizability to the within-tissue performances on the GDSC and PRISM datasets, while statistically significant F1 performance improvements on the CCLE (8%, p-value=0.001) and DrugComb (19%, p-value=0.009) datasets. Evaluation on the challenging blind tests suggests GPT's competitiveness on the CCLE and DrugComb datasets compared to random splitting. Furthermore, analyses of the drug-pathway associations and log probabilities provided valuable insights that align with previous DSP findings.

Conclusion

The diverse experiment setups and in-depth analysis underscore the importance of generative LLM, such as GPT, as a viable in silico approach to guide precision oncology.

Availability

https://github.com/bioIKEA/SensitiveCancerGPT.

Elucidating the role of artificial intelligence in drug development from the perspective of drug-target interactions

Drug development remains a critical issue in the field of biomedicine. With the rapid advancement of information technologies such as artificial intelligence (AI) and the advent of the big data era, AI-assisted drug development has become a new trend, particularly in predicting drug-target associations. To address the challenge of drug-target prediction, AI-driven models have emerged as powerful tools, offering innovative solutions by effectively extracting features from complex biological data, accurately modeling molecular interactions, and precisely predicting potential drug-target outcomes. Traditional machine learning (ML), network-based, and advanced deep learning architectures such as convolutional neural networks (CNNs), graph convolutional networks (GCNs), and transformers play a pivotal role. This review systematically compiles and evaluates AI algorithms for drug- and drug combination-target predictions, highlighting their theoretical frameworks, strengths, and limitations. CNNs effectively identify spatial patterns and molecular features critical for drug-target interactions. GCNs provide deep insights into molecular interactions via relational data, whereas transformers increase prediction accuracy by capturing complex dependencies within biological sequences. Network-based models offer a systematic perspective by integrating diverse data sources, and traditional ML efficiently handles large datasets to improve overall predictive accuracy. Collectively, these AI-driven methods are transforming drug-target predictions and advancing the development of personalized therapy. This review summarizes the application of AI in drug development, particularly in drug-target prediction, and offers recommendations on models and algorithms for researchers engaged in biomedical research. It also provides typical cases to better illustrate how AI can further accelerate development in the fields of biomedicine and drug discovery.

Network Medicine as a Tool to Enhance Regenerative Rehabilitation Practice

OBJECTIVE

The aim of the study was to demonstrate the utility of network medicine for enhancing regenerative rehabilitation approaches.

DESIGN

We employed a scientometric approach to illustrate the scientific evolution of regenerative rehabilitation and network medicine over the past 20 yrs. We then present two exemplars of a novel application of network medicine, first to optimize the development of a multimodal rehabilitation program for osteoarthritis and, second, to demonstrate the potential utility of a regenerative rehabilitation protocol for neurological applications.

RESULTS

The scientometric analysis revealed thematic clusters guiding the fields of regenerative rehabilitation, while also revealing a notable gap in the integration of advanced computational biology tools, such as network medicine. To address this gap, we applied a network medicine paradigm as a proof-of-principle in rehabilitation, demonstrating its ability to predict previously known combinational benefits of a multimodal rehabilitation protocol for osteoarthritis and potential of a regenerative rehabilitation protocol to enhance outcomes after stroke.

CONCLUSIONS

The trajectory of progress in regenerative rehabilitation has been defined by embracing new technologies and approaches. We propose that network medicine is a new frontier for the field. Toward the design of efficient and optimized clinical protocols, we anticipate that the network medicine paradigm may have broad applications in the field of rehabilitation, even beyond the examples presented here.

Fusing Micro- and Macro-Scale Information to Predict Anticancer Synergistic Drug Combinations

Drug combination therapy is highly regarded in cancer treatment. Computational methods offer a time- and cost-effective opportunity to explore the vast combination space. Although deep learning-based prediction methods lead the field, their generalization ability remains unsatisfactory. Few previous studies have the ability to finely characterize drugs and cell lines at both the micro-scale and macro-scale. Furthermore, the interaction of cross-scale information is often overlooked. These two points limit models' ability of predicting the synergism of drug combinations in cell lines. To address the issues, we propose a novel anticancer synergistic drug combination prediction method termed MMFSynergy in this article. The construction of MMFSynergy involves three phases. First, MMFSynergy pretrains two micro encoders and a macro graph encoder, which can capture micro- or macro-scale information from large volumes of unlabeled data and generate generic features for drugs and proteins. Second, it represents drugs and proteins by fusing cross-scale information through a self-supervised task. Finally, it employs a Transformer Encoder-based model to predict synergy scores, taking representations of drugs in the combinations and the associated proteins of cell lines as input. We compared our method with eight advanced methods across three typical scenarios based on two public datasets. The results consistently demonstrated that the proposed method's generalization ability outperforms six advanced methods'. We also conducted experiments including but not limited to ablation study and case study to further exhibit the effectiveness of MMFSynergy.

Advancing personalized cancer therapy: Onko_DrugCombScreen-a novel Shiny app for precision drug combination screening

Identifying and validating genotype-guided drug combinations for a specific molecular subtype in cancer therapy represents an unmet medical need and is important in enhancing efficacy and reducing toxicity. However, the exponential increase in combinatorial possibilities constrains the ability to identify and validate effective drug combinations. In this context, we have developed Onko_DrugCombScreen, an innovative tool aiming at advancing precision medicine based on identifying significant drug combination candidates in a target cancer cohort compared to a comparison cohort. Onko_DrugCombScreen, inspired by the molecular tumor board process, synergizes drug knowledgebase analysis with various statistical methodologies and data visualization techniques to pinpoint drug combination candidates. Validated through a TCGA-BRCA case study, Onko_DrugCombScreen has demonstrated its proficiency in discerning established drug combinations in a specific cancer type and in revealing potential novel drug combinations. By enhancing the capability of drug combination discovery through drug knowledgebases, Onko_DrugCombScreen represents a significant advancement in personalized cancer treatment by identifying promising drug combinations, setting the stage for the development of more precise and potent combination treatments in cancer care. The Onko_DrugCombScreen Shiny app is available at https://rshiny.gwdg.de/apps/onko_drugcombscreen/. The Git repository can be accessed at https://gitlab.gwdg.de/MedBioinf/mtb/onko_drugcombscreen.

Targeting disease: Computational approaches for drug target identification

With the advancing technology, the way to drug discovery has evolved. The use of AI and computational methods have revolutionized the methods to develop novel therapeutics. In previous years, the methods to discover new drugs included high-throughput screening and bioassays which were labor-dependent, extremely expensive and had high probability to inaccurate results. The introduction of Computational studies has changed the process by introducing various methods to determine hit compounds and their methods of analysis. Methods such as molecular docking, virtual screening, and dynamics have changed the path to optimize and produce lead molecules. Similarly, network pharmacology also works on the identification of target proteins complex disease pathways with the help of protein-protein interactions and obtaining hub proteins. Various tools such as STRING database, cytoscape and metascape are employed in the study to construct a network between the proteins responsible for the disease progression and helps to obtain the vital target proteins, simplifying the process of drug-target identification. These approaches when employed together, results in obtaining results with better precision and accuracy which can be further validated experimentally, saving the resources and time. This chapter highlights the foundation of computational approaches in drug discovery and provides a detailed understanding of how these approaches are helping the researchers to produce novel solutions using artificial intelligence and machine learning.

Identifying traditional Chinese medicine combinations for breast cancer treatment based on transcriptional regulation and chemical structure

Breast cancer (BC) is a prevalent form of cancer among women. Despite the emergence of numerous therapies over the past few decades, few have achieved the ideal therapeutic effect due to the heterogeneity of BC. Drug combination therapy is seen as a promising approach to cancer treatment. Traditional Chinese medicine (TCM), known for its multicomponent nature, has been validated for its anticancer properties, likely due to the synergy effect of the key components. However, identifying effective component combinations from TCM is challenging due to the vast combination possibilities and limited prior knowledge. This study aims to present a strategy for discovering synergistic compounds based on transcriptional regulation and chemical structure. First, BC-related gene sets were used to screen TCM-derived compound combinations guided by synergistic regulation. Then, machine learning models incorporating chemical structural features were established to identify potential compound combinations. Subsequently, the pair of honokiol and neochlorogenic acid was selected by integrating the results of compound combination screening. Finally, cell experiments were conducted to confirm the synergistic effect of the pair against BC. Overall, this study offers an integrated screening strategy to discover compound combinations of TCM against BC. The tumor cell suppression effect of the honokiol and neochlorogenic acid pair validated the effectiveness of the proposed strategy.

Simplicity within biological complexity

Motivation

Heterogeneous, interconnected, systems-level, molecular (multi-omic) data have become increasingly available and key in precision medicine. We need to utilize them to better stratify patients into risk groups, discover new biomarkers and targets, repurpose known and discover new drugs to personalize medical treatment. Existing methodologies are limited and a paradigm shift is needed to achieve quantitative and qualitative breakthroughs.

Results

In this perspective paper, we survey the literature and argue for the development of a comprehensive, general framework for embedding of multi-scale molecular network data that would enable their explainable exploitation in precision medicine in linear time. Network embedding methods (also called graph representation learning) map nodes to points in low-dimensional space, so that proximity in the learned space reflects the network's topology-function relationships. They have recently achieved unprecedented performance on hard problems of utilizing few omic data in various biomedical applications. However, research thus far has been limited to special variants of the problems and data, with the performance depending on the underlying topology-function network biology hypotheses, the biomedical applications, and evaluation metrics. The availability of multi-omic data, modern graph embedding paradigms and compute power call for a creation and training of efficient, explainable and controllable models, having no potentially dangerous, unexpected behaviour, that make a qualitative breakthrough. We propose to develop a general, comprehensive embedding framework for multi-omic network data, from models to efficient and scalable software implementation, and to apply it to biomedical informatics, focusing on precision medicine and personalized drug discovery. It will lead to a paradigm shift in the computational and biomedical understanding of data and diseases that will open up ways to solve some of the major bottlenecks in precision medicine and other domains.

Revealing rutaecarpine's promise: A pathway to parkinson's disease relief through PPAR modulation

The pathological mechanisms of Parkinson's disease (PD) is complex, and no definitive cure currently exists. This study identified Rutaecarpine (Rut), an alkaloid extracted from natural plants, as a potential therapeutic agent for PD. To elucidate its mechanisms of action and specific effects in PD, network pharmacology, molecular docking, and experimental validation methods were employed. Our findings demonstrated the efficacy of Rut in ameliorating PD symptoms. Network pharmacology analysis indicated that Rut exerts its therapeutic effects through the PPAR signaling pathway and the lipid pathway. Molecular docking results revealed that Rut forms stable protein-ligand complexes with PPARα and PPARγ. Animal experiments showed that Rut improved motor function in PD mice, protected dopaminergic neurons, ameliorated lipid metabolism disorders, and reduced neuroinflammation. This study identified the critical molecular mechanisms and therapeutic targets of Rut in the treatment of PD, providing a theoretical foundation for future investigations into the pharmacodynamics of Rut as a potential anti-PD agent.

Personalized prediction of anticancer potential of non-oncology drugs through learning from genome derived molecular pathways

Advances in cancer genomics have significantly expanded our understanding of cancer biology. However, the high cost of drug development limits our ability to translate this knowledge into precise treatments. Approved non-oncology drugs, comprising a large repository of chemical entities, offer a promising avenue for repurposing in cancer therapy. Herein we present CHANCE, a supervised machine learning model designed to predict the anticancer activities of non-oncology drugs for specific patients by simultaneously considering personalized coding and non-coding mutations. Utilizing protein-protein interaction networks, CHANCE harmonizes multilevel mutation annotations and integrates pharmacological information across different drugs into a single model. We systematically benchmarked the performance of CHANCE and show its predictions are better than previous model and highly interpretable. Applying CHANCE to approximately 5000 cancer samples indicated that >30% might respond to at least one non-oncology drug, with 11% non-oncology drugs predicted to have anticancer activities. Moreover, CHANCE predictions suggested an association between SMAD7 mutations and aspirin treatment response. Experimental validation using tumor cells derived from seven patients with pancreatic or esophageal cancer confirmed the potential anticancer activity of at least one non-oncology drug for five of these patients. To summarize, CHANCE offers a personalized and interpretable approach, serving as a valuable tool for mining non-oncology drugs in the precision oncology era.

BiRNN-DDI: A Drug-Drug Interaction Event Type Prediction Model Based on Bidirectional Recurrent Neural Network and Graph2Seq Representation

Research on drug-drug interaction (DDI) prediction, particularly in identifying DDI event types, is crucial for understanding adverse drug reactions and drug combinations. This work introduces a Bidirectional Recurrent Neural Network model for DDI event type prediction (BiRNN-DDI), which simultaneously considers structural relationships and contextual information. Our BiRNN-DDI model constructs drug feature graphs to mine structural relationships. For contextual information, it transforms drug graphs into sequences and employs a two-channel structure, integrating BiRNN, to obtain contextual representations of drug-drug pairs. The model's effectiveness is demonstrated through comparisons with state-of-the-art models on two DDI event-type benchmarks. Extensive experimental results reveal that BiRNN-DDI surpasses other models in accuracy, AUPR, AUC, F1 score, Precision, and Recall metrics on both small and large datasets. Additionally, our model exhibits a lower parameter space, indicating more efficient learning of drug feature representations and prediction of potential DDI event types.

Enhancing drug discovery in schizophrenia: a deep learning approach for accurate drug-target interaction prediction - DrugSchizoNet

Drug discovery relies on the precise prognosis of drug-target interactions (DTI). Due to their ability to learn from raw data, deep learning (DL) methods have displayed outstanding performance over traditional approaches. However, challenges such as imbalanced data, noise, poor generalization, high cost, and time-consuming processes hinder progress in this field. To overcome the above challenges, we propose a DL-based model termed DrugSchizoNet for drug interaction (DI) prediction of Schizophrenia. Our model leverages drug-related data from the DrugBank and repoDB databases, employing three key preprocessing techniques. First, data cleaning eliminates duplicate or incomplete entries to ensure data integrity. Next, normalization is performed to enhance security and reduce costs associated with data acquisition. Finally, feature extraction is applied to improve the quality of input data. The three layers of the DrugSchizoNet model are the input, hidden and output layers. In the hidden layer, we employ dropout regularization to mitigate overfitting and improve generalization. The fully connected (FC) layer extracts relevant features, while the LSTM layer captures the sequential nature of DIs. In the output layer, our model provides confidence scores for potential DIs. To optimize the prediction accuracy, we utilize hyperparameter tuning through OB-MOA optimization. Experimental results demonstrate that DrugSchizoNet achieves a superior accuracy of 98.70%. The existing models, including CNN-RNN, DANN, CKA-MKL, DGAN, and CNN, across various evaluation metrics such as accuracy, recall, specificity, precision, F1 score, AUPR, and AUROC are compared with the proposed model. By effectively addressing the challenges of imbalanced data, noise, poor generalization, high cost and time-consuming processes, DrugSchizoNet offers a promising approach for accurate DTI prediction in Schizophrenia. Its superior performance demonstrates the potential of DL in advancing drug discovery and development processes.

DDINet: Drug-drug interaction prediction network based on multi-molecular fingerprint features and multi-head attention centered weighted autoencoder

Drug-drug interactions (DDIs) pose a major concern in polypharmacy due to their potential to cause unexpected side effects that can adversely affect a patient's health. Therefore, it is crucial to identify DDIs effectively during the early stages of drug discovery and development. In this paper, a novel DDI prediction network (DDINet) is proposed to enhance the predictive performance over conventional DDI methods. Leveraging the DrugBank dataset, drugs are represented using the Simplified Molecular Input Line-Entry System (SMILES), with the RDKit software pre-processing the SMILES strings into their canonical forms. Multiple molecular fingerprinting techniques such as Extended Connectivity Fingerprints (ECFPs), Molecular ACCess System keys (MACCSkeys), PubChem Fingerprints, 3D molecular fingerprints (3D-FP), and molecular dynamics fingerprints (MDFPs) are employed to encode drug chemical structures into feature vectors. Drug similarities are computed using the Tanimoto coefficient (TC), and the final Structural Similarity Profile (SSP) is obtained by averaging the five molecular fingerprint types. The novelty of the approach lies in the integration of a Multi-head Attention centered Weighted Autoencoder (Mul_WAE) as the interaction prediction module, which leverages the Multi-head Attention (MHA) layer to focus on the most significant input features. Furthermore, we introduce the Upgraded Bald Eagle Search Optimization (UBesO) algorithm, which optimally selects the learnable parameters of the Mul_WAE based on cross-entropy loss, improving the model's convergence and performance. The proposed DDINet model achieves an accuracy of 99.77%, 99.66% of AUC, 99.5% average precision, 99.4% precision, and 99.49% recall, providing a comprehensive evaluation of the model's robustness. Beyond high accuracy, DDINet offers advantages in scalability, making it well suited for handling large datasets due to its efficient feature extraction and optimization processes. The unique combination of multiple molecular fingerprinting methods with the MHA layer and UBesO algorithm highlights the innovative aspects of our model and significantly improves prediction performance compared to existing approaches.

Comparable clinical advantages identification of three formulae on rheumatic disease using a modular-based network proximity approach

ETHNOPHARMACOLOGICAL RELEVANCE

Herbal formulae have been used in China for thousands of years but have unclear clinical positioning and unknown characteristic indications make it difficult to determine their specific application in various diseases, which seriously hamper their clinical value. Identifying the precise clinical positioning and clinical advantages of similar formulae for related diseases is a critical issue.

AIM OF THIS STUDY

To develop a methodology based on modular pharmacology to determine the clinical advantages and precise clinical position of similar formulae.

MATERIALS AND METHODS

In this study, we proposed a modular-based network proximity approach to explore drug repositioning and clinical advantages of three formulae, Shirebi tablets (SRB), Yuxuebi capsules (YXB), and Wangbifukang granules (WBFK), for rheumatic disease. First, we constructed a rheumatology target network, and modules were obtained using the cluster tool molecular complex detection (MCODE). Based on the modular interaction map established by a quantitative approach for inter-module coordination and its transition (IMCC), using a targeting rate (TR) matrix to identify targeted modules of three formulae. Moreover, the network proximity Z-score and Jaccard similarity coefficient were used to identify potential optimal symptomatic indications and related diseases using three formulae. At the same time, the driver genes for SRB and gouty arthritis were identified by flow centrality and shortest distance, and the epresentative driver genes were validated by in vivo experiments.

RESULTS

32 modules were obtained using the MCODE method. 4, 4, and 14 characteristic targeted modules of SRB, YXB, and WBFK, respectively, were identified using a targeting rate (TR) matrix. Module 2, 16, and 19 were targeted by both SRB and WBFK. The common effects of SRB and WBFK focused on inflammatory response and innate immune response, YXB was found to be involved in the collagen catabolic process, transmembrane receptor protein serine/threonine kinase signaling pathway. Moreover, potential optimal symptomatic indications and representative related diseases were identified for three formulae: SRB was significantly associated with GA (Z = -20.26); YXB was significantly associated with AS (Z = -4.532), MI (Z = -29.11), RhFv (Z = -6.945), OA (Z = -39.97), and GA (Z = -13.03); and WBFK was significantly associated with MI (Z = -205.5), SLE (Z = -37.65), RhFv (Z = -42.45), and GA (Z = -17.24). Finally, 8 driver genes for SRB and gouty arthritis were identified,the representative driver genes TRAF6 and NFE2L1 were validated by in vivo experiments.

CONCLUSIONS

The modular-based network proximity approach proposed in this study may provide a new perspective for the precise drug repositioning and clinical advantages of similar formulae in disease treatment.

Network-Based Identification of Key Toxic Compounds in Airborne Chemical Exposome

Air pollution is a leading contributor to the global disease burden. However, the complex nature of the chemicals to which humans are exposed through inhalation has obscured the identification of the key compounds responsible for diseases. Here, we develop a network topology-based framework to identify key toxic compounds in the airborne chemical exposome. Using cardiovascular diseases (CVDs) as a model disease, we found that toxic network modules of various compounds are closely linked to the modules of CVDs. The proximity of compound target modules to disease modules can indicate the extent of toxicity induced by the compounds. By integrating mass spectrometry-based external exposure concentrations and machine learning-predicted internal exposure concentrations, we established a comprehensive linkage connecting exposure to disease-related risk for the identification of toxic compounds. These findings were subsequently validated using exposure and disease data on the regional scale. This work provides an effective strategy for identifying key compounds within environmental exposomes and establishes a new paradigm for understanding the pathogenicity of air pollution.

Optimizing kinase and PARP inhibitor combinations through machine learning and in silico approaches for targeted brain cancer therapy

The drug combination is an attractive approach for cancer treatment. PARP and kinase inhibitors have recently been explored against cancer cells, but their combination has not been investigated comprehensively. In this study, we used various drug combination databases to build ML models for drug combinations against brain cancer cells. Some decision tree-based models were used for this purpose. The results were further evaluated using molecular docking and molecular dynamics (MD) simulation. The possibility of the hit drug combinations for crossing the Blood-brain barrier (BBB) was also examined. Based on the obtained results, the combination of niraparib, as the PARP inhibitor, and lapatinib, as the kinase inhibitor, exhibited more considerable outcomes with a remarkable model performance (accuracy of 0.915) and prediction confidence of 0.92. The protein tweety homolog 3 and BTB/POZ domain-containing protein 2 are the main targets of niraparib and lapatinib with - 10.2 and - 8.5 scores, respectively. Due to the outcomes, this drug combination can use the CAT1 transporter on the BBB surface and effectively cross the BBB. Based on the obtained results, niraparib-lapatinib can be a promising drug combination candidate for brain cancer treatment. This combination is worth to be examined by experimental investigation in vitro and in vivo.

Integrating pharmacogenomics and cheminformatics with diverse disease phenotypes for cell type-guided drug discovery

BACKGROUND

Large-scale pharmacogenomic resources, such as the Connectivity Map (CMap), have greatly assisted computational drug discovery. However, despite their widespread use, CMap-based methods have thus far been agnostic to the biological activity of drugs as well as to the genomic effects of drugs in multiple disease contexts. Here, we present a network-based statistical approach, Pathopticon, that uses CMap to build cell type-specific gene-drug perturbation networks and integrates these networks with cheminformatic data and diverse disease phenotypes to prioritize drugs in a cell type-dependent manner.

METHODS

We build cell type-specific gene-drug perturbation networks from CMap data using a statistical procedure we call Quantile-based Instance Z-score Consensus (QUIZ-C). Using these networks and a large-scale disease-gene network consisting of 569 disease signatures from the Enrichr database, we calculate Pathophenotypic Congruity Scores (PACOS) between input gene signatures and drug perturbation signatures and combine these scores with cheminformatic data from ChEMBL to prioritize drugs. We benchmark our approach by calculating area under the receiver operating characteristic curves (AUROC) for 73 gene sets from the Molecular Signatures Database (MSigDB) using target gene expression profiles from the Comparative Toxicogenomics Database (CTD). We validate the drugs predicted in our proofs-of-concept using real-time polymerase chain reaction (qPCR) experiments.

RESULTS

Cell type-specific gene-drug perturbation networks built using QUIZ-C are topologically distinct, reflecting the biological uniqueness of the cell lines in CMap, and are enriched in known drug targets. Pathopticon demonstrates a better prediction performance than solely cheminformatic measures as well as state-of-the-art network and deep learning-based methods. Top predictions made by Pathopticon have high chemical structural diversity, suggesting their potential for building compound libraries. In proof-of-concept applications on vascular diseases, we demonstrate that Pathopticon helps guide in vitro experiments by identifying pathways that are potentially regulated by the predicted therapeutic candidates.

CONCLUSIONS

Our network-based analytical framework integrating pharmacogenomics and cheminformatics (available at https://github.com/r-duh/Pathopticon ) provides a feasible blueprint for a cell type-specific drug discovery and repositioning platform with broad implications for the efficiency and success of drug development.

DrugMAP 2.0: molecular atlas and pharma-information of all drugs

The escalating costs and high failure rates have decelerated the pace of drug development, which amplifies the research interests in developing combinatorial/repurposed drugs and understanding off-target adverse drug reaction (ADR). In other words, it is demanded to delineate the molecular atlas and pharma-information for the combinatorial/repurposed drugs and off-target interactions. However, such invaluable data were inadequately covered by existing databases. In this study, a major update was thus conducted to the DrugMAP, which accumulated (a) 20831 combinatorial drugs and their interacting atlas involving 1583 pharmacologically important molecules; (b) 842 repurposed drugs and their interacting atlas with 795 molecules; (c) 3260 off-targets relevant to the ADRs of 2731 drugs and (d) various types of pharmaceutical information, including diverse ADMET properties, versatile diseases, and various ADRs/off-targets. With the growing demands for discovering combinatorial/repurposed therapies and the rapidly emerging interest in AI-based drug discovery, DrugMAP was highly expected to act as an indispensable supplement to existing databases facilitating drug discovery, which was accessible at: https://idrblab.org/drugmap/.

Angelic acid triggers ferroptosis in colorectal cancer cells via targeting and impairing NRF2 protein stability

Ferroptosis is a unique programmed cell death driven by iron-dependent phospholipid peroxidation. Tumor cells that escape from the conventional therapies appear more sensitive to ferroptosis. Therefore, it is extremely urgent to find safe and efficient active ingredients that induce ferroptosis in tumor cells. Herein, we identified that angelic acid, as a potent anti-tumor active ingredient in Angelica sinensis, profoundly sensitizes CRC cells to ferroptosis via a natural compound library screen. We revealed that angelic acid treatment is sufficient to predispose CRC cells to ferroptosis phenotype, evidenced by malondialdehyde (MDA) accumulation, lipid peroxidation, and upregulation of ferroptosis-associated markers CHAC1 and PTGS2, which is abolished by ferroptosis inhibitor Fer-1. Moreover, the results of network pharmacology showed that NRF2 is critical for angelic acid-mediated CRC cell ferroptosis. Mechanistically, angelic acid exerted its ferroptosis induction via directly binding and facilitating the degradation of NRF2. In addition, the syngeneic mouse models revealed that angelic acid boosts CRC cell sensitivity to ferroptosis inducer sulfasalazine with essentially no toxicity in vivo. Collectively, our findings highlighted a previously unrecognized anti-tumor mechanism of angelic acid and represented an appealing therapeutic strategy for CRC treatment.

Genetic Commonalities Between Metabolic Syndrome and Rheumatic Diseases Through Disease Interactome Modules

This study aims to elucidate the potential genetic commonalities between metabolic syndrome (MetS) and rheumatic diseases through a disease interactome network, according to publicly available large-scale genome-wide association studies (GWAS). The analysis included linkage disequilibrium score regression analysis, cross trait meta-analysis and colocalisation analysis to identify common genetic overlap. Using modular partitioning, the network-based association between the two disease proteins in the protein-protein interaction set was divided and quantified. Clinical samples from public databases were used to confirm the mapped genes. Mendelian randomisation analyses were conducted using genetic instrumental variables for causal inference. MetS and rheumatoid arthritis (RA), ankylosing spondylitis (AS), systemic lupus erythematosus (SLE), Sjogren's syndrome (SS) and their primary module networks shared topological overlap and genetic correlation. Functional analysis highlighted the significance of these shared targets in processes such as a diverse array of metabolic pathways involving glucose, lipids, energy, protein transport, inflammatory response, autophagy and cytokine regulation, elucidating the pathways through which MetS intersects with rheumatic diseases. Causal associations were determined between MetS phenotypes and rheumatic diseases. The persistence of MetS effects on rheumatic diseases remained evident even after adjusting for alcohol consumption and smoking. We have highlighted specific genetic associations between MetS and rheumatic diseases. Several genes (e.g., PRRC2A, PSMB8, BAG6, GPSM3, PBX2, etc.) have been identified with molecular commonalities in MetS and RA, AS, SLE and SS, which may serve as potential targets for shared treatments.

Integrating Ultra-High-Performance Liquid Chromatography and Orbitrap High-Resolution Mass Spectrometry, Feature-Based Molecular Networking, and Network Medicine to Unlock Harvesting Strategies for Endangered Sinocalycanthus Chinensis

Evaluating the practical utility of endangered plant species is crucial for their conservation. Nevertheless, numerous endangered plants, including Sinocalycanthus chinensis, lack historical usage data, leading to a paucity of guidance in traditional pharmacological research. This gap impedes their development and potential utilization. Ultra-high-performance liquid chromatography and Orbitrap high-resolution mass spectrometry were employed to analyze the S. chinensis leaves collected at different harvesting times. Then, the metabolites were automatically annotated by a self-built R script in conjunction with characteristic fragment ions, neutral loss filtering, and feature-based molecular networking. By integrating metabolomics with network medicine analysis, the potential usage and optimal harvest times for S. chinensis were unlocked. A total of 305 metabolites were identified, with 66.8% annotated by self-built R script. A progressive increase in metabolite disparities was observed from May to August, followed by a relatively minor distinction from August to October. Notably diverse metabolites were detected in S. chinensis harvested during different periods, implying potential variations in efficacy. Network medicine analysis indicated possible therapeutic implications of S. chinensis for lung cancer, diabetes, bladder cancer, and Alzheimer's disease. Samples collected in May and September demonstrated exceptional efficacy. Harvesting was strategically conducted during these months based on variations in sample characteristics and metabolite content, tailored to their intended applications for dietary or medicinal purposes. This study developed an efficient methodology for investigating metabolites and exploring the potential applications of S. chinensis in food and herbal medicine. Consequently, it provides technical support for the sustainable conservation of endangered plants with limited clinical application experience.

Predicting the toxic side effects of drug interactions using chemical structures and protein sequences

The study aims to address the critical issue of toxic side effects resulting from drug combinations, which can significantly increase health risks, clinical complications, and lead to drug being withdrawn from the market. A model named TSEDDI (toxic side effects of drug-drug interaction) has been developed to improve the identification of drug pairs that may induce toxicity or adverse reactions. By utilizing drug chemical structures and diverse proteins, we employ a convolutional neural network (CNN) to extract features from molecular images, enzyme proteins, transporter proteins, and target proteins. Furthermore, we introduce a weighted binary cross entropy loss function to tackle class imbalance and integrate the multi-head attention mechanism with residual connections to enhance model performance. Our model outperformed advanced baseline models in predicting drug-drug interaction (DDI) side effects, achieving an accuracy of 0.9059 (± 0.0010) and consistently excelling across various evaluation metrics. The case study confirms the potential mechanisms by which four pairs of drugs cause side effects, thus demonstrating the effectiveness of our model in predicting DDI side effects. The TSEDDI model combines multiple attention mechanisms and residual connections, enhancing its ability to detect toxic and adverse effects related to DDIs. As a result, it becomes a valuable resource for promptly identifying adverse reactions in clinical trials. Future research could investigate drug substructures prone to toxic side effects.

Simulate Scientific Reasoning with Multiple Large Language Models: An Application to Alzheimer's Disease Combinatorial Therapy

Motivation

This study aims to develop an AI-driven framework that leverages large language models (LLMs) to simulate scientific reasoning and peer review to predict efficacious combinatorial therapy when data-driven prediction is infeasible.

Results

Our proposed framework achieved a significantly higher accuracy (0.74) than traditional knowledge-based prediction (0.52). An ablation study highlighted the importance of high quality few-shot examples, external knowledge integration, self-consistency, and review within the framework. The external validation with private experimental data yielded an accuracy of 0.82, further confirming the framework's ability to generate high-quality hypotheses in biological inference tasks. Our framework offers an automated knowledge-driven hypothesis generation approach when data-driven prediction is not a viable option.

Availability and implementation

Our source code and data are available at https://github.com/QidiXu96/Coated-LLM.

Network-based artificial intelligence approaches for advancing personalized psychiatry

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

Advancing Anti-Biofilm Strategies: Innovations to Combat Biofilm-Related Challenges and Enhance Efficacy

Biofilms are complex communities of microorganisms that can cause significant challenges in various settings, including industrial processes, environmental systems, and human health. The protective nature of biofilms makes them resistant to traditional anti-biofilm strategies, such as chemical agents, mechanical interventions, and surface modifications. To address the limitations of conventional anti-biofilm methods, researchers have explored emerging strategies that encompass the use of natural compounds, nanotechnology-based methods, quorum-sensing inhibition, enzymatic degradation, and antimicrobial photodynamic/sonodynamic therapy. There is an increasing focus on combining multiple anti-biofilm strategies to combat resistance and enhance effectiveness. Researchers are continuously investigating the mechanisms of biofilm formation and developing innovative approaches to overcome the limitations of conventional anti-biofilm methods. These efforts aim to improve the management of biofilms and prevent infections while preserving the environment. This study provides a comprehensive overview of the latest advancements in anti-biofilm strategies. Given the dynamic nature of this field, exploring new approaches is essential to stimulate further research and development initiatives. The effective management of biofilms is crucial for maintaining the health of industrial processes, environmental systems, and human populations.