Artificial intelligence (AI) has emerged as a powerful tool for solving real world problems across a wide range of industries and is increasingly being utilised by pharmaceutical companies to discover novel drug targets, biomarkers, and new drugs. Several AI-driven small molecules have entered clinical trials over the past few years, but their fate remains unknown. Currently, no commercially available compounds have been developed solely using AI approaches. In this perspective, we examine the current use of AI in drug discovery for neuropsychiatry. The pace of drug discovery in neuropsychiatric medicine has been generally sluggish, largely due to challenges such as poor pharmacological selectivity, the blood-brain barrier, and a limited understanding of disease mechanisms. AI may offer innovative solutions to these challenges. However, relative to fields such as oncology, the impact of AI on the discovery of neuropsychiatric drugs has been limited. Although novel AI-tools have been developed to overcome some of the challenges involved in neuropsychiatry drug discovery, their effectiveness has not been sufficiently evaluated. To date, innovative tools such as AlphaFold have been used to identify drug candidates for multiple neuropsychiatric conditions. AI-driven platforms have been used to study behavioural data from preclinical models to identify novel clinical candidates in clinical trials (e.g., ulotaront, phase III). It is anticipated that the availability of large-scale multi-omics data ('big data') will likely increase in the future, allowing us to gain a better understanding of gene-associated mechanisms in psychiatry. Using AI-based technologies such as AlphaFold, future pharmacological targets will be identified based on gene expression data, and large libraries of chemical compounds will be screened rapidly to identify novel drug candidates, resulting in shorter pre-clinical phase with lower costs.
Natural products have inspired the discovery of several drug candidates and US Food and Drug Administration-approved drugs. However, the conventional drug development pathway has several limitations, necessitating innovative strategies. Computational pharmacology and in silico clinical trials (ISCTs) have emerged as a potential to optimize the drug discovery process. Current literature rarely discusses these 2 domains together, leaving limited guidance on how computational tools used in natural product research can complement ISCT frameworks. This review addresses this gap by summarizing key computational methods used for natural products and highlighting how their outputs can inform ISCT-based evaluation of efficacy, safety, and translational potential. Computational approaches such as molecular docking, pharmacophore modeling, quantitative structure-activity relationship analysis, absorption-distribution-metabolism-excretion-toxicity prediction, and molecular dynamics simulations have been applied in natural products-based drug discovery. The findings of these studies are promising and suggest favorable prospects for the use of computational methods in identifying, optimizing, and evaluating bioactive natural compounds. Notably, this approach minimizes the attrition rate in later clinical stages, as the most promising compounds with auspicious pharmacokinetic and pharmacodynamic profiles are chosen. ISCTs further hold considerable value in drug discovery through investigating the safety and efficacy of investigational drugs using simulated or virtual patient populations, ensuring a lower attrition rate when wet-lab clinical trials are conducted. Although the ISCTs on natural products are limited, the integration of computational pharmacology and ISCTs shows potential in the discovery of valuable natural compounds and the acceleration of drug development from natural products. Despite existing challenges in data quality and regulatory acceptance, the integration of emerging technologies such as machine learning, artificial intelligence, and hybrid platforms shows promise in advancing natural product-based therapeutics and precision medicine. SIGNIFICANCE STATEMENT: This review brings together computational pharmacology and in silico clinical trials to address persistent challenges in natural product drug development. It highlights how early computational screening, absorption-distribution-metabolism-excretion-toxicity prediction, and dynamic modeling can inform virtual clinical evaluation, reduce attrition, and accelerate translation. The paper offers a clear and practical roadmap for improving efficiency, strengthening safety assessment, and advancing precision in the development of natural product-based drugs.
Natural products (NPs) have historically yielded numerous therapeutic agents, yet their integration into modern drug discovery has been constrained by chemical complexity, low abundance, laborious dereplication, and limited target annotation. Convergence of multi-omics technologies with high-resolution structural and biological data has created unprecedented opportunities for artificial intelligence (AI) to accelerate NP-based therapeutics development. This review provides an operational, end-to-end workflow that explicitly connects computational predictions to medicinal chemistry decision points, addressing a critical gap between computational prediction and clinical translation. We trace the complete discovery pipeline: computational mining of biosynthetic gene clusters (BGCs) and metabolomes, deep learning (DL)-assisted structural elucidation and dereplication, network-based target identification using protein-ligand prediction, and generative molecular design inspired by NP scaffolds (including large language models, diffusion models, and genetic algorithms). Critical evaluation of current limitations (data scarcity, lack of standardized ontologies, model interpretability) is complemented by discussion of emergent strategies (foundation models trained on multi-modal data, graph neural networks, autonomous closed-loop laboratories). Representative case studies, including the synthetic AI-designed clinical benchmark rentosertib, illustrate the current evidence spectrum from discovery-level validation to early clinical benchmarking, while also highlighting that most AI-enabled NP discovery workflows remain at the preclinical or proof-of-concept stage, with limited quantitative evidence of improved clinical productivity. We conclude with an Outlook proposing feasible developments for 2025-2030: self-driving laboratories with reported acceleration in specific experimental contexts, foundation models enabling hypothesis-free chemical space exploration, and sustainability-aware AI frameworks embedding biodiversity impact assessments. This operational focus fills a critical gap between algorithmic capability and clinically actionable NP-derived leads. Importantly, while AI has demonstrably accelerated several early discovery steps, quantitative comparisons with classical NP workflows remain limited, and most reported advances are supported by preclinical or proof-of-concept studies rather than systematic evidence of improved time-to-lead, cost reduction, or clinical success rates.
Neurological disorders refer to a diverse group of conditions that affect the brain, peripheral nerves, and spinal cord and impair socioemotional, cognitive, motor, and sensory functions. Alzheimer's disease (AD), Multiple Sclerosis (MS), Parkinson's disease (PD), Huntington's disease (HD), and Amyotrophic Lateral Sclerosis (ALS) are some of the well-known neurodegenerative diseases that affect millions of people worldwide. Despite the advanced technologies and nano-drug delivery systems, the success rate of developing drugs for neurological disorders is significantly low. Among several constraints, including gastrointestinal irritation, rapid metabolism, and low stability, the blood-brain barrier (BBB) emerges as one of the key challenges in the development and application of drugs against neurological disorders. These challenges necessitate innovative approaches to develop cost-effective therapeutic strategies. Drug repurposing, the discovery of new therapeutic benefits of existing drugs, is a promising drug discovery approach for discovering potential treatment options for complex neurological disorders. This review aims to explore the advanced and significant progress in drug repurposing for major neurological disorders, including MS, AD, PD, ALS, HD, stroke, and neuropsychiatric conditions. It places an explicit emphasis on discussing the potential role of artificial intelligence (AI)-assisted drug repurposing and understanding of the biological mechanisms in discovering new drugs for these neurological conditions. This also examines current challenges in drug repurposing and provides a critical review of the available opportunities and limitations in AI-assisted drug repurposing.
Fibrotic diseases represent a major global health burden, with current therapies often limited by modest efficacy and substantial side effects. Traditional Chinese medicine (TCM) provides a rich yet underexplored reservoir for anti-fibrotic drug discovery, valued for its chemical diversity and extensive clinical history. Nonetheless, conventional approaches to deriving therapeutics from these multi-component formulas face significant challenges, including low tissue selectivity, limited target specificity, suboptimal scaffold optimization, and incomplete cross-scale validation. Here, we summarize recent advances in discovering TCM-derived anti-fibrotic compounds and highlight emerging strategies, including tissue-oriented active compound discovery, fibrosis-target mining, scaffold optimization guided by targets and biosynthesis, and cross-scale validation. Building on these developments, we introduce the Smart Herbal-based Innovation and Translational Engine (SHINE), an intelligent discovery platform that prioritizes promising herbal candidates using modern computational and experimental methods; innovative development of optimized compounds or novel derivatives; a translational pipeline to unite preclinical findings and clinical validation; and a closed-loop feedback system where real-world evidence continuously informs drug repositioning and secondary development. SHINE bridges TCM principles with state-of-the-art drug discovery technologies, enabling the development of well-defined, mechanism-based small-molecule candidates from classical formulas. By integrating empirical herbal knowledge with advanced multi-omics, artificial intelligence, and biosynthetic engineering, SHINE aims to deliver first-in-class anti-fibrotic therapeutics with defined targets, improved safety, and demonstrable clinical efficacy.
The emergence of structurally complex therapeutic modalities, including bispecific antibodies, antibody-drug conjugates, fusion proteins, incretins, radioligand therapeutics, antibody-oligonucleotide conjugates, and small interfering RNAs, demands advanced mass spectrometry workflows across the ADME pipeline. This forward-looking perspective examines the bioanalytical challenges these next-generation therapeutics present, from absorption and distribution through metabolism and elimination. Current MS platforms, LC-MS/MS, high-resolution MS, quantitative mass spectrometry imaging, and ICP-MS, while highly capable, present opportunities for sensitivity enhancement to fully characterize low-dose therapeutics through the elimination phase. To quantify this need, we developed a pharmacokinetic-driven prediction framework and applied it to FDA-approved complex therapeutics, demonstrating that enhanced sensitivity would enable comprehensive metabolite characterization, particularly during the elimination phase. We identify five convergent innovation priorities: (1) addressing modality complexity through integrated ADME workflows; (2) continued advancement of instrument sensitivity for low-dose therapeutics; (3) resolving biotransformation challenges through enhanced analytical resolution; (4) improving qMSI capabilities for therapeutic-level tissue detection; and (5) implementing AI/ML-driven automation for data complexity management. Rather than advocating for MS-only solutions, we recommend integration of mass spectrometric structural specificity with complementary high-throughput technologies; immunoassays, hybridization ELISA, qPCR, and element-specific detection, to enable comprehensive bioanalysis across all modalities, accelerating the path from discovery to development. Modern medicines are becoming more complex. New treatments such as antibody–drug conjugates, RNA-based therapies, and bispecific antibodies are made of large, intricate molecules that behave very differently from traditional drugs inside the body. To develop these treatments safely, scientists need to understand how the body breaks them down and what breakdown products are formed, a process known as ADME characterization (Absorption, Distribution, Metabolism, and Elimination). Mass spectrometry, which measures the mass of molecules with high precision, is a powerful tool for this work. However, as treatments become more complex, their breakdown products often appear at extremely low concentrations in blood and tissue, sometimes far below what current instruments can reliably detect. This perspective examines the challenges of characterizing these next-generation therapeutics, including instrument sensitivity limitations, resolving complex molecular breakdown patterns, and implementing artificial intelligence to manage data complexity. We highlight the urgent need for more sensitive mass spectrometry technologies and recommend integrated analytical strategies combining multiple measurement platforms to close this “sensitivity gap” and accelerate the development of next-generation therapeutics.
Enterovirus 71 (EV71) is a major pathogen causing severe and fatal hand, foot, and mouth disease. Its strong neurotropism and rapid evolution pose an ongoing threat to infants and young children. To date, no specific antiviral drug against EV71 has been approved, leaving a critical gap in clinical management. This review summarizes recent progress in the search for anti-EV71 drugs. The structural features of EV71 and key steps in its life cycle, including entry and replication, are first described. The evolution of drug discovery technologies is then traced, from traditional cytopathic effect-based screening to modern platforms such as computer-assisted virtual screening, reporter-carrying pseudovirus systems, and high-content imaging. Next, two main antiviral strategies are discussed: direct-acting agents that target the viral capsid, protease, and polymerase, and host-targeted approaches that modulate virus-dependent pathways including metabolism, signal transduction, and immune responses. In addition, the review highlights recent medicinal chemistry efforts in structure-based optimization of EV71 inhibitors, covering side-chain modifications, cyclization, prodrug design, and covalent binding engineering. By summarizing the mechanisms, optimization strategies, and current state of candidate drugs, this review aims to provide a practical reference for the development of antivirals against EV71 and other enteroviruses.
Differences among individuals in susceptibility to adverse drug reactions are strongly influenced by genetic variation affecting drug metabolism, transport, immune response, and pharmacodynamics pathways, presenting both challenges and opportunities for drug discovery and development. Genetic and genomic variation can modify pharmacokinetics, pharmacodynamics, immunemediated toxicity, treatment resistance, and biomarker-defined therapeutic benefit, making pharmacogenetics and pharmacogenomics central components of precision medicine. The primary aim of this review is to synthesize how pharmacogenetics and pharmacogenomics function as enabling technologies across the pharmaceutical and clinical domains, with emphasis on biomarker-guided drug development, patient stratification, safety optimization, regulatory translation, precision oncology, sex-related variability, population diversity, artificial intelligence, and multi-omics integration. The review also elucidates the distinction between single-gene pharmacogenetics and genome-wide pharmacogenomics and discusses the practical requirements, procedures, advantages, and challenges associated with implementation. Future progress will depend on harmonized standards, diverse population datasets, preemptive testing linked to electronic health records, explainable artificial intelligence, multi-omics validation, and equitable access to genomic prescribing tools, ensuring that precision therapeutics improve drug safety and efficacy across populations.
Precise control over how and where small-molecule drugs are covalently attached to monoclonal antibodies is increasingly vital for creating consistently efficacious and safe antibody-drug conjugates (ADCs) as powerful targeted therapies. Enzymatic conjugation methods have gained considerable attention for their abilities to efficiently install payloads at defined locations under mild and predictable conditions. This work provides a comprehensive review of current enzymatic technologies for site-specific ADC construction, organized by the biological origin of enzymes. Among various enzyme-based conjugating approaches, a special focus is given to the emerging ADP-ribosyl cyclase-enabled ADC (ARC-ADC) platform. By utilizing genetically fused CD38, a member of the ARC family, together with its dinucleotide-derived inhibitor, site-specific ADCs with defined drug-to-antibody ratios in varied formats could be facilely produced with demonstrated efficacy and specificity in preclinical models of different types of cancer. Unlike most enzymatic methods requiring recognition tags or external catalytic steps, ARC-ADC provides a fully integrated, modular strategy for streamlined ADC discovery and development.
Autoimmune hepatitis (AIH) is an immune-mediated chronic liver disease with an increasing incidence. The complex pathogenesis of AIH poses significant challenges for clinical treatment. In recent years, with the introduction of cutting-edge concepts such as "immune-metabolic interplay" and "intercellular communication networks," along with novel detection methods and technologies, research on the pathogenesis of AIH has shifted from a single-molecular-mechanism perspective to a systematic regulatory network analysis. This shift has not only provided more research evidence for a deeper understanding of the molecular biological mechanisms underlying AIH but also offers a potential reference for the development of targeted therapeutic drugs for AIH based on novel targets. Therefore, this review starts with aberrant activation signals from key immune cells-including dendritic cells (DCs), macrophages (Mφ), and T/B cells-and integrates the intercellular signaling communication mechanisms between hepatocytes, cholangiocytes, and immune cells to systematically summarize the key molecular biological mechanisms and targets identified in recent years, providing a reference for future elucidation of the critical mechanisms of AIH. On this basis, the review further integrates the current application and research progress of clinically used AIH therapeutic drugs, as well as those at various stages of development, including potential therapeutic compounds. It discusses the limitations of current clinical drugs and evaluates the feasibility and future application potential of potential compounds for AIH treatment in preclinical and clinical studies, thereby offering comprehensive research evidence for the management of AIH.
The COVID-19 pandemic highlighted the role of rapid viral mutation and global connectivity in accelerating viral emergence and spread, emphasising the necessity for expedited and adaptable antiviral drug discovery and development. Despite ongoing efforts to develop effective, low-toxicity therapeutics, the number of antivirals that have achieved clinical approval remains limited. The shortfall is especially significant in developing countries, where access to new antivirals is limited by high prices, few options, import dependence, unstable supply chains and weak purchasing systems. The challenge is further heightened by the emergence of increasingly drug-resistant variants while vaccines often provide inadequate protection against newly mutated or novel viruses. Consequently, the identification of novel antiviral agents that are both effective and cost-efficient via innovative strategies for antiviral drug discovery is essential to manage and control viral infections. Therefore, this review examines the different challenges associated with conventional antiviral drugs alongside recent strategies in antiviral drug discovery and development, such as the exploitation of plant secondary metabolites with antiviral properties, advanced microscopy technologies, computer-aided drug design, artificial intelligence and machine learning, gene-editing technologies, drug combination therapy and nanotechnology-enhanced drug delivery systems. Additionally, this study proposes a simple decision-focused pathway integrating natural products, computation and targeted delivery to guide candidate prioritisation, optimisation and translation from discovery to implementation. While these emerging strategies offer considerable promise, challenges related to validation, toxicity, scalability and equitable access remain important considerations for successful clinical translation. Future research should therefore integrate complementary technologies to accelerate the development of effective antiviral agents against current and emerging viral threats.
Respiratory diseases, encompassing chronic inflammatory conditions, interstitial fibrotic disorders, acute infectious diseases, and pulmonary malignancies, represent a profound global health burden with unacceptably high morbidity and mortality rates. Historically, the pharmaceutical pipeline for respiratory therapeutics has suffered staggering attrition rates during clinical development. This is primarily due to the fundamental inability of conventional two-dimensional cell cultures and in vivo animal models to faithfully recapitulate the complex three-dimensional architecture, multicellular heterogeneity, and human-specific physiological dynamics of the pulmonary system. To bridge this critical translational gap, lung organoids-self-organizing, three-dimensional microphysiological constructs derived from pluripotent or adult stem cells-have emerged as a useful human-cell-based platform. This comprehensive review critically evaluates current lung organoid technologies, elucidating their derivation pathways and capacity for high-fidelity disease modeling. We analyze their application in dissecting the pathogenesis of chronic obstructive pulmonary disease, idiopathic pulmonary fibrosis, viral infections including SARS-CoV-2, and non-small cell lung cancer. Furthermore, we highlight their important role in predictive toxicology for assessing environmental inhalation hazards, cosmetic safety, and drug-induced lung injury, aligning with evolving regulations prioritizing alternatives to animal testing. Despite their immense potential, widespread clinical and industrial translation is currently impeded by biological bottlenecks: the absence of functional vascularization, incomplete immune integration, and reliance on undefined xenogeneic matrices. We systematically examine bioengineering strategies addressing these limitations-including synthetic hydrogels, microfluidic organ-on-a-chip platforms, and 3D bioprinting-to overcome translational hurdles, accelerate precision medicine, and improve respiratory pharmacology.
Multiple Myeloma (MM) is a hematological malignancy characterized by the clonal proliferation and survival of neoplastic plasma cells (PCs) within the bone marrow (BM), where disease progression is critically supported by interactions with the BM tumor microenvironment (TME). Despite significant advances in therapeutic strategies, MM remains incurable, underscoring the need for improved preclinical models to better understand the disease biology and therapeutic response. This review summarizes current and emerging MM treatment approaches and critically examines the development of models designed to more accurately recapitulate interactions between MM-PCs and the surrounding BM niche. We describe established and emerging modeling platforms, with emphasis on advanced three-dimensional (3D) culture systems and highlight their unique contributions to the preclinical assessment of both existing and novel therapies. The advantages of 3D models, including in vitro and in silico systems, over traditional two-dimensional (2D) models are discussed, alongside a comparative evaluation of scaffold-free and scaffold-based approaches. In addition, the benefits and recent advances in the customization of BM niche simulation using microfluidic technologies and organ-on-a-chip platforms are reviewed. The application of 3D models in MM research is increasingly enabling the study of disease pathogenesis, progression, drug resistance and precision-medicine approaches (informed by biomarker discovery). Although standardized preclinical approaches for evaluating MM therapeutics are currently lacking, the growing imperative to reduce reliance on preclinical animal models highlights the importance of alternate systems. Consequently, the development and adoption of physiologically relevant models that accurately recapitulate MM-PC interactions with the BM TME will be critical for advancing future therapeutic strategies in MM.
Acinetobacter baumannii, a Gram-negative member of the ESKAPE pathogens, has emerged as a pivotal cause of multidrug-resistant hospital-acquired or nosocomial infections worldwide. Its ability to regulate virulence and biofilm formation through quorum sensing (QS) significantly contributes to its virulence and pathogenicity. Acyl-homoserine lactone synthase (AHLS) from the A. baumannii AYE strain plays a key role in the QS pathway and represents a promising druggable target for the development of anti-bacterial strategies. A homology-modeled three-dimensional structure of AHLS (AYE strain) was predicted, optimized, and validated. High-throughput virtual screening of 975 natural antimicrobial compounds was performed, followed by Lipinski's and ADMET profiling to assess drug-likeness and safety. Promising drug candidates were further evaluated using 100ns molecular dynamics (MD) simulations to identify putative AHLS inhibitors. MM/PBSA based binding free energy calculations revealed favorable interactions for CID_291096 (-14.74 ± 2.20 kcal/mol), CID_155586 (-15.26 ± 2.27 kcal/mol), and MSID_001127 (-28.44 ± 3.32 kcal/mol). Among these, MSID_001127 (Lovastatin) demonstrated superior structural stability and sustained intermolecular non-covalent interactions throughout the 100ns MD simulation. Structural stability was further supported by RMSD, RMSF, Rg, SASA, PCA, and hydrogen-bonding analyses. Through virtual screening, three phytochemical lead compounds targeting AHLs with high negative binding free energies were identified. Stable protein-ligand interactions and favourable binding energetics were identified by molecular docking, 100 ns molecular dynamics simulations, and MM/PBSA analyses. Based on the results, ligand MSID_001127 was the most promising lead candidate compared with cipargamin. However, additional experimental validation is required to verify its therapeutic potential and biological activity. These findings suggest that Lovastatin may be a promising drug candidate for AHLS targeting the QS pathway of A. baumannii. The results warrant further experimental validation to explore its potential as an anti-bacterial therapeutic agent.
Food allergy (FA) is an immune-mediated adverse reaction to food components (mainly proteins) and represents an increasing global health problem, affecting millions of individuals and imposing substantial clinical, psychosocial, and economic burdens. Accurate diagnosis is essential to prevent life-threatening reactions while avoiding unnecessary dietary restrictions and impaired quality of life. Current diagnostic approaches rely on clinical history, skin prick tests (SPT), measurement of serum allergen-specific IgE (sIgE), and oral food challenges (OFC). However, SPT and sIgE are highly sensitive but lack specificity, frequently identifying clinically irrelevant sensitization, whereas OFC remains the diagnostic gold standard despite being resource-demanding and carrying a substantial risk of systemic reactions. In recent years, several innovative diagnostic approaches have emerged with the aim of improving diagnostic accuracy and reducing reliance on OFC. Component-resolved diagnostics (CRD) enable detailed characterization of molecular sensitization profiles, supporting improved risk stratification and identification of clinically relevant cross-reactivity patterns. Functional cellular assays, including the basophil activation test (BAT) and the mast cell activation test (MAT), offer direct assessment of IgE-mediated effector cell responses and have demonstrated higher diagnostic specificity compared with conventional testing. Epitope-specific IgE profiling and the allergen-specific IgG4/IgE ratio may additionally contribute to a better understanding of disease phenotype and evolution. Furthermore, advances in multi-omics technologies combined with artificial intelligence and machine learning are creating new opportunities for biomarker discovery and predictive modelling in FA. This narrative review summarizes current innovative diagnostic techniques in food allergy, discussing their clinical applications, limitations, and future directions toward more precise and personalized diagnostic approaches.
Microbiology laboratories play a critical role in the diagnosis and management of infectious diseases. However, recent advancements aimed at reducing human workload and minimizing time loss are gaining popularity. Artificial intelligence (AI) technologies, particularly machine learning (ML) and deep learning (DL), have been reported to contribute significantly to microbial laboratory diagnostics. Through this approach, molecular methods, genetic sequencing, microbiological meta-analyses, and related fields benefit from faster and more accurate analytic capabilities. In addition to diagnostic applications, AI is increasingly used in genomics, metagenomics, antimicrobial resistance (AMR) prediction, and drug and vaccine discovery, enabling more comprehensive and data-driven microbiological analysis. This review comprehensively evaluates current AI applications in microbiology, highlighting their advantages, limitations, and implementation challenges. It further examines the suitability of different AI methodologies for specific laboratory tasks and compares AI-driven approaches with conventional expert-based practices. Finally, the study emphasizes the complementary roles of AI systems and human expertise, underscoring their synergistic potential to improve diagnostic accuracy, efficiency, and clinical decision-making.
Liver fibrosis is the common consequence of liver injury caused by a variety of chronic liver disorders. This condition leads to the development of more severe complications, particularly cirrhosis and hepatocellular carcinoma. Despite abundant studies, the fundamental cell and molecular mechanisms of liver fibrosis are still unknown. There are many key players involved in the initiation and progression of liver fibrosis. Thus, the specific type of underlying disease and the study's objectives should be considered while choosing suitable models for liver fibrosis. Numerous in vitro and in vivo models have been generated to investigate liver fibrosis and proposed for drug screening and toxicology; however, there are no ideal in vitro models for drug discovery yet. In this review, we introduced the available in vitro models and highlighted certain platforms such as organoids and liver-on-a-chip for investigating liver fibrosis. Furthermore, we discussed the current challenges and potential application of each model.
Secondary metabolites derived from medicinal plants represent a critical resource for drug discovery and play an indispensable role in the prevention and treatment of major human diseases. Spatial metabolomics has emerged as a transformative approach that overcomes the inherent limitations of traditional metabolomics, particularly tissue homogenisation, by enabling precise, in situ profiling of the spatial heterogeneity and dynamic accumulation of key bioactive compounds. When integrated with complementary technologies such as spatial multi-omics (SMO), single-cell analysis, and artificial intelligence (AI)-assisted imaging, this paradigm provides a powerful framework for deciphering the spatiotemporal regulatory networks that govern secondary metabolite biosynthesis. This review systematically compares conventional and spatial metabolomics, synthesises recent advances in medicinal plant research, critically evaluates current technical challenges and optimisation strategies, and outlines future directions enabled by multi-omics integration and intelligent computational analytics. Collectively, these insights aim to establish a theoretical foundation for the in-depth investigation of medicinal plants and to support their translational application and industrial advancement.
HIV-1 infects CD4+ T-cells, causing immunodeficiency and, if untreated, progression to AIDS. Global antiretroviral therapy scale-up has reduced deaths and new infections by nearly half over the past decade, yet lifelong treatment and emerging drug resistance remain critical challenges. This review provides an overview of the HIV-1 lifecycle, with emphasis on the structure, function, and multiple roles of IN in viral replication. We discuss the development, mechanisms of action, clinical utility, and resistance profiles of current integrase inhibitors, underscoring the need for innovative approaches to drug discovery. Emerging strategies that exploit non-catalytic functions of integrase are critically evaluated as promising avenues for overcoming resistance and expanding the antiviral repertoire. The biochemical, biophysical, and cell-based screening platforms that have enabled integrase inhibitor discovery and continue to drive next-generation drug development are examined as a critical component of the integrase drug discovery ecosystem. Recent advances in assay technologies, structural biology, and computational approaches are highlighted for their potential to expand the integrase-targeting therapeutic arsenal and address the growing challenge of antiviral resistance.
The review is devoted to the use of artificial intelligence (AI) in scientific research and development to create new or repurpose authorized drug products, as well as to the use of AI-based solutions to discover new biomarkers and shorten the time to diagnosis of various diseases. The following areas of AI application are considered: the search for new pharmacologically active substances, the development of formulations and drug production technology, preclinical trials, and intelligent diagnostics (identification of new biomarkers; development of software products to interpret research results and increase diagnostic accuracy). Examples of AI use by leading pharmaceutical companies and a list of the most popular AI models in drug development are provided. The revolutionary contribution of AI in drug discovery lies in reducing the time to identify new drug candidate molecules by more rapidly identifying potential biotargets, performing virtual screening, optimizing promising candidates based on predictive data on pharmacokinetic and toxicological profiles, and searching for the optimal way to synthesize potential drugs. In addition, another area of AI application is the development of drug-delivery devices and systems that improve patient compliance and usability. This paper presents examples of AI use in intelligent diagnostics that prove their high accuracy and time efficiency compared to conventional methods of diagnostics, risk assessment, and prognosis in oncology, cardiology, and other areas of medicine. The implementation of AI technologies in medicine is intensifying, raising questions of ethics, the quality and adequacy of data, the effectiveness and safety of results for patients, personnel competence and readiness for change, as well as issues related to intellectual property rights.