Search Results (357)

The discovery and development of new drugs is a lengthy, complex, and costly process, often requiring 10–20 years to progress from initial concept to market approval, with clinical trials representing the most resource-intensive stage. In recent years, Artificial Intelligence (AI) has emerged as a transformative technology capable of reshaping the entire pharmaceutical research and development (R&D) pipeline. The purpose of this narrative review is to examine the role of AI in drug discovery and development, highlighting its contributions, challenges, and future implications for pharmaceutical sciences and global public health. A comprehensive review of the scientific literature was conducted, focusing on published studies, reviews, and reports addressing the application of AI across the stages of drug discovery, preclinical development, clinical trials, and post-marketing surveillance. Key themes were identified, including AI-driven target identification, molecular screening, de novo drug design, predictive toxicity modelling, and clinical monitoring. The reviewed evidence indicates that AI has significantly accelerated drug discovery and development by reducing timeframes, costs, and failure rates. AI-based approaches have enhanced the efficiency of target identification, optimized lead compound selection, improved safety predictions, and supported adaptive clinical trial designs. Collectively, these advances position AI as a catalyst for innovation, particularly in promoting accessible, efficient, and sustainable healthcare solutions. However, substantial challenges remain, including reliance on high-quality and representative biomedical data, limited algorithmic transparency, high implementation costs, regulatory uncertainty, and ethical and legal concerns related to data privacy, bias, and equitable access. In conclusion, AI represents a paradigm shift in pharmaceutical research and drug development, offering unprecedented opportunities to improve efficiency and innovation. Addressing its technical, ethical, and regulatory limitations will be essential to fully realize its potential as a sustainable and globally impactful tool for therapeutic innovation. Full article

X-ray crystallography remains the gold standard for high-resolution structural biology, yet obtaining diffraction-quality crystals continues to pose a major bottleneck due to inherently low success rates. This review advocates a paradigm shift from probabilistic screening to rational engineering, reframing crystallization as a controllable self-assembly process. We provide a comprehensive overview of strategies that connect fundamental physicochemical principles to practical applications, beginning with contact design, which involves the active engineering of crystal contacts through surface entropy reduction (SER), introduction of electrostatic patches. Complementing these molecular approaches, we discuss physicochemical strategies that exploit heterogeneous nucleation on functionalized surfaces and gold nanoparticles (AuNPs) to lower the energy barrier for crystal formation. We also address scaffold design, utilizing rigid fusion partners and polymer-forming chaperones to promote crystallization even from low-concentration solutions. Furthermore, we highlight principles for controlling the behavior of multi-component complexes, based on our experimental experience. Finally, we examine de novo lattice design, which leverages AI tools such as AlphaFold and RFdiffusion to program crystal lattices from first principles. Together, these strategies establish an integrated workflow that links thermodynamic stability with crystallizability. Full article

Background: The emergence of therapy-related myelodysplastic syndrome (t-MN) after autologous stem cell transplantation (ASCT) is well documented. However, with the growing use of chimeric antigen receptor (CAR) T-cell therapy for relapsed/refractory B-cell malignancies, concerns about secondary myeloid neoplasms, particularly MN, have arisen. The mechanisms and cytogenetic features associated with post-CAR-T MN, especially chromosome 7 abnormalities, remain underexplored. Objectives: To compare the incidence, timing, and cytogenetic characteristics of MN developing after CAR-T-cell therapy versus ASCT, and to evaluate the potential association between CAR-T therapy, persistent cytopenias, and these specific alterations. Study Design: This was a retrospective, single-center study of 275 patients with B-cell malignancies treated between 2015 and 2024 at Hospital Universitario Central de Asturias (Spain). Of these, 259 patients underwent ASCT and 16 received CAR-T-cell therapy (axicabtageneciloleucel n = 13, tisagenlecleucel n = 2, brexucabtageneautoleucel n = 1). Clinical, cytogenetic, and laboratory data were collected and analyzed. Incidence rates were compared using Fisher’s exact test, and time-to-event outcomes was evaluated using the Mann–Whitney U test (given the small number of events). Statistical significance was set at p < 0.05. Results: Myeloid neoplasms were diagnosed in 3 of 259 ASCT patients (1.15%) and in 2 of 16 CAR-T-cell patients (12.5%) (p = 0.03). The median time to myeloid neoplasm diagnosis was numerically shorter in the CAR-T group (15.5 vs. 69 months, p = 0.096). All post-CAR-T cases presented persistent cytopenias and cytokine release syndrome (CRS). Cytogenetic analyses revealed de novo monosomy 7 and 7q deletion in both CAR-T-related cases, whereas no chromosome 7 abnormalities were detected in ASCT-related cases. Pre-treatment samples did not show these abnormalities, although limitations in the sensitivity of the assays preclude the definitive exclusion of minor pre-existing clones. Both affected CAR-T patients had prolonged CAR-T cell persistence and required transfusional support due to hematologic toxicity. One patient was diagnosed with high-risk MN with 5q and 7q deletion and the other with Clonal Cytopenia of Uncertain Significance (CCUS) with monosomy 7. Conclusions: CAR-T-cell therapy was associated with a significantly higher and earlier incidence of myeloid neoplasms compared to ASCT in this cohort. The development of post-CAR-T myeloid neoplasm was characterized by persistent cytopenias, prolonged CAR-T cell persistence, and de novo chromosome 7 alterations. While the small sample size necessitates cautious interpretation, these findings may suggest a distinct pathogenesis potentially linked to inflammation, immune toxicity, or the expansion of pre-existing clones. This highlights the need for long-term hematologic monitoring and evaluation for clonal hematopoiesis prior to CAR-T-cell therapy, especially in heavily pretreated patients. Full article

Enzyme engineering drives innovation in biotechnology, medicine, and industry, yet conventional approaches remain limited by labour-intensive workflows, high costs, and narrow sequence diversity. Artificial intelligence (AI) is revolutionising this field by enabling rapid, precise, and data-driven enzyme design. Machine learning and deep learning models such as AlphaFold2, RoseTTAFold, ProGen, and ESM-2 accurately predict enzyme structure, stability, and catalytic function, facilitating rational mutagenesis and optimisation. Generative models, including ProteinGAN and variational autoencoders, enable de novo sequence creation with customised activity, while reinforcement learning enhances mutation selection and functional prediction. Hybrid AI–experimental workflows combine predictive modelling with high-throughput screening, accelerating discovery and reducing experimental demand. These strategies have led to the development of synthetic “synzymes” capable of catalysing non-natural reactions, broadening applications in pharmaceuticals, biofuels, and environmental remediation. The integration of AI-based retrosynthesis and pathway modelling further advances metabolic and process optimisation. Together, these innovations signify a shift from empirical, trial-and-error methods to predictive, computationally guided design. The novelty of this work lies in presenting a unified synthesis of emerging AI methodologies that collectively define the next generation of enzyme engineering, enabling the creation of sustainable, efficient, and functionally versatile biocatalysts. Full article

Background: Pregnancy-associated breast cancer (PABC), defined as breast cancer diagnosed during pregnancy or within one year postpartum, is a unique and clinically challenging entity. Evidence suggests that tumors diagnosed during pregnancy (PrBC) and postpartum (PPBC) may differ in biology and prognosis. This study compares clinical features, treatment patterns and outcomes between PrBC and PPBC. Methods: We performed a retrospective analysis of 76 women diagnosed with PABC from January 2000 to June 2023 across two tertiary centers. Patients were classified according to ESMO guidelines as PrBC (n = 41) or PPBC (n = 35). Clinical presentation, tumor characteristics, treatment approaches and survival outcomes were evaluated. Overall survival (OS) and disease-free survival (DFS) were estimated using Kaplan–Meier analysis and compared with log-rank tests. Results: A total of 76 patients with PABC were included (41 PrBC, 35 PPBC; median age 37 years). Most tumors were high-grade invasive ductal carcinomas, with Luminal B predominant in PrBC and triple-negative breast cancer (TNBC) in PPBC. Locally advanced disease was common (axillary involvement 52%; de novo metastases 9%). Surgery was performed in most cases, with breast conservative surgery (BCS) more frequent in PrBC and mastectomy in PPBC; 46% received neoadjuvant chemotherapy. At median follow-up of 68 months, 7.9% of patients had died and 29% experienced recurrence. Oncologic outcomes were similar between subgroups, with a trend in favor of PrBC. Pregnancy continuation did not adversely affect outcomes. Conclusions: PrBC and PPBC display heterogeneous clinical presentations with a trend toward more favorable outcomes in PrBC. These findings support the need for tailored counseling, individualized management and research designs that differentiate between PrBC and PPBC. Full article

Antimicrobial resistance (AMR) has become a major health crisis worldwide, and it is expected to surpass cancer as one of the leading causes of death by 2050. Conventional antibiotics are struggling to keep pace with the rapidly evolving resistance trends, underscoring the urgent need for novel antimicrobial therapeutic strategies. Antimicrobial peptides (AMPs) function through diverse, often membrane-disrupting mechanisms that can address the latest challenges to resistance. However, the identification, prediction, and optimization of novel AMPs can be impeded by several issues, including extensive sequence spaces, context-dependent activity, and the higher costs associated with wet laboratory screenings. Recent developments in artificial intelligence (AI) have enabled large-scale mining of genomes, metagenomes, and quantitative species-resolved activity prediction, i.e., MIC, and de novo AMPs designed with integrated stability and toxicity filters. The current review has synthesized and highlighted progress across different discriminative models, such as classical machine learning and deep learning models and transformer embeddings, alongside graphs and geometric encoders, structure-guided and multi-modal hybrid learning approaches, closed-loop generative methods, and large language models (LLMs) predicted frameworks. This review compares models’ benchmark performances, highlighting AI-predicted novel hybrid approaches for designing AMPs, validated by in vitro and in vivo methods against clinical and resistant pathogens to increase overall experimental hit rates. Based on observations, multimodal paradigm strategies are proposed, focusing on identification, prediction, and characterization, followed by design frameworks, linking active-learning lab cycles, mechanistic interpretability, curated data resources, and uncertainty estimation. Therefore, for reproducible benchmarks and interoperable data, collaborative computational and wet lab experimental validations must be required to accelerate AI-driven novel AMP discovery to combat multidrug-resistant Gram-negative pathogens. Full article

Helichrysum italicum (Roth) G. Don (Asteraceae) is a valuable medicinal and aromatic plant native to a variety of habitats across the Mediterranean region. However, genetic studies of this morphologically diverse species have been limited by the scarcity of species-specific DNA markers. To address this limitation, we generated the first de novo transcriptome assembly comprising 24,806 transcripts from young shoots containing leaves and flowers, developed EST-SSR markers, and evaluated their utility in population genetic analysis. Seventy-eight primer pairs were designed, of which 23 showed successful amplification, polymorphism, and transferability to Helichrysum litoreum Guss. and Helichrysum arenarium (L.) Moench. A subset of 12 EST-SSRs was used to genotype 270 individuals from 12 natural populations of H. italicum in Corsica (France), along with one outgroup population from Croatia. The polymorphic information content ranged from 0.250 to 0.796, and Shannon’s information index ranged from 0.588 to 1.843, indicating the markers’ suitability for population genetic studies. Analysis of molecular variance revealed that 15% of the total genetic variation was attributable to differences among populations. Discriminant analysis of principal components and Bayesian clustering in STRUCTURE identified distinct population clusters corresponding to geographic locations. Notably, the southernmost coastal populations were clearly differentiated from the others. Full article

Plant microRNAs (miRNAs) are endogenous non-coding RNAs (~20–24 nucleotides) that regulate gene expression post-transcriptionally, playing critical roles in plant growth, development, and stress responses. This review systematically examines AI applications in plant miRNA research, tracing evolution from traditional machine learning to deep learning architectures. Plant miRNAs exhibit distinctive features necessitating plant-specific computational approaches: nuclear-localized biogenesis, high target complementarity (>80%), and coding region targeting. These characteristics enable more accurate computational prediction and experimental validation than animal systems. Methodological advances have improved prediction accuracy from ~90% (early SVMs) to >99% (recent deep learning), though metrics reflect different evaluation contexts. We analyze applications across miRNA identification, target prediction with degradome validation, miRNA–lncRNA interactions, and ceRNA networks. Critical assessment reveals that degradome data capture mixed RNA fragments from multiple sources beyond miRNA cleavage, requiring stringent multi-evidence validation. Similarly, fundamental ambiguities in lncRNA definition compound prediction uncertainties. Major challenges include severe data imbalance (positive to negative ratios of 1:100 to 1:10,000), limited cross-species generalization, insufficient model interpretability, and experimental validation bottlenecks. Approximately 75% of plant miRNA families in miRBase v20 lack convincing evidence, underscoring the need for rigorous annotation standards. Future directions encompass multimodal deep learning, explainable AI, spatiotemporal graph neural networks, and ultimately AI-driven de novo miRNA design, though the latter requires substantial advances in both computation and high-throughput validation. This synthesis demonstrates that AI has become indispensable for plant miRNA research, providing essential support for crop improvement while acknowledging persistent challenges demanding continued innovation. Full article

The genetic improvement of Phalaenopsis-type Dendrobium, a valuable ornamental and medicinal orchid, is hindered by the lack of a complete reference genome. In this study, a transcriptome-based approach was employed to develop and validate insertion–deletion (InDel) markers for genetic analysis and variety identification. RNA-seq was performed on two distinct varieties, resulting in the de novo assembly of 156,108 unigenes. A bioinformatics pipeline was developed to identify 5083 high-quality InDel loci, from which 1029 potential markers were designed. Fifty primer pairs were selected and validated experimentally, with 84% successfully amplifying clear products, and 76% showing polymorphism. The polymorphism information content (PIC) of the markers ranged from 0.25 to 0.78, indicating their high potential for use in genetic diversity studies. These markers were used to classify 24 Phalaenopsis-type Dendrobium varieties into distinct genetic clusters. This work provides a scalable and robust platform for molecular breeding, DNA fingerprinting, and germplasm management in non-model species that lack a reference genome. By leveraging transcriptome data, these markers will contribute to the efficient genetic improvement of Dendrobium and other similar crops. Full article

Background/Objectives: GuaB, which is known as inosine 5′-phosphate dehydrogenase (IMPDH), is an enzymatic target involved in the de novo guanine biosynthetic pathway of the multidrug-resistant (MDR) Acinetobacter baumannii. GuaB has emerged as a potential therapeutic target to cope with increasing antibiotic resistance. Here, we used machine learning-based virtual screening as a verification technique to find potential inhibitors possessing different chemical scaffolds, using structure-based drug design as a discovery platform. Methods: Four machine learning models, built based on chemical fingerprint data, were trained, and the best models were used for virtual screening of the ChEMBL library, which covers 153 active molecules. Molecular dynamics (MD) simulations of 200 ns were carried out for all three compounds in order to explain conformational changes, evaluate stability, and provide validation of the docking results. Post-simulation analyses include principal component analysis (PCA), bond analysis, free-energy landscape (FEL), dynamic cross-correlation matrix (DCCM), radial distribution function (RDF), salt-bridge identification, and secondary-structure profiling, etc. Results: For molecular docking, the screened compounds were used against the GuaB protein to achieve proper docked conformation. Upon visual examination of the best-docked compounds, three leads (lead-1, lead-2, and lead-3) were found to have better interaction with the GuaB protein in comparison to the control. The mean RMSD scores between the three leads and the control were between 2.54 and 2.89 Å. In addition, the three leads as well as the control were characterized for pharmacokinetic features. All three leads met Lipinski’s Rule 5 and were thus drug-like. PCA and FEL analyses showed that lead-2 exhibited improved conformational stability, identified as deeper energy minima, whereas RDF and DCCM analyses revealed that lead-2 and lead-3 exhibited strong local structuring and concerted dynamics. In addition, lead-2 displayed a very rich hydrogen-bonding network with a total of 460 frames possessing such interactions, which is the highest among the complexes investigated here. Based on entropy calculations and the maximum entropy method of gamma–gram, lead-1 proved to be the most stable one with the lowest binding free-energy. Conclusions: This study provides an integrated machine learning-based virtual screening pipeline for the identification of new scaffolds to moderate infections associated with AMR; however, in vitro validation is still required to assess the efficacy of such compounds. Full article

Colorectal cancer is one of the most dangerous neoplastic diseases in terms of both mortality and incidence. Thus, anti-colorectal cancer agents are urgently needed. Computational approaches have great potential to accelerate the phenotypic discovery of versatile anticancer agents. Here, by combining perturbation-theory machine learning (PTML) modeling with the fragment-based topological design (FBTD) approach, we provide key computational evidence on the computer-aided de novo design and prediction of new molecules virtually exhibiting multi-cell inhibitory activity against different colorectal cancer cell lines. The PTML model created in this study achieved sensitivity and specificity values exceeding 80% in training and test sets. The FBTD approach was employed to physicochemically and structurally interpret the PTML model. These interpretations enabled the rational design of six new drug-like molecules, which were predicted as active against multiple colorectal cancer cell lines by both our PTML model and a CLC-Pred 2.0 webserver, with the latter being a well-established virtual screening tool for early anticancer discovery. This work confirms the potential of the joint use of PTML and FBTD as a unified computational methodology for early phenotypic anticancer drug discovery. Full article

Multi-target drug design represents a paradigm shift in tackling the complexity and heterogeneity of diseases such as cancer. Conventional single-target therapies frequently face limitations due to network redundancy, pathway compensation, and adaptive resistance mechanisms. In contrast, deep generative models, empowered by advanced artificial intelligence algorithms, provide scalable and versatile platforms for the de novo generation and optimization of small molecules with activity across multiple therapeutic targets. This review provides a comprehensive overview of the recent landscape of AI-driven deep generative modeling for multi-target drug discovery, highlighting breakthroughs in model architectures, molecular representations, and goal-directed optimization strategies. We also examine the emergence of self-improving learning systems, closed-loop frameworks that iteratively refine molecular candidates through integrated feedback, as a transformative approach to adaptive drug design. Finally, key challenges, current limitations, and emerging trends are discussed to guide the evolution of next-generation intelligent and autonomous drug discovery pipelines for multi-target therapeutics. Full article

Diffusion models have emerged as a leading framework in generative modeling, poised to transform the traditionally slow and costly process of drug discovery. This review provides a systematic comparison of their application in designing two principal therapeutic modalities: small molecules and therapeutic peptides. We dissect how the unified framework of iterative denoising is adapted to the distinct molecular representations, chemical spaces, and design objectives of each modality. For small molecules, these models excel at structure-based design, generating novel, pocket-fitting ligands with desired physicochemical properties, yet face the critical hurdle of ensuring chemical synthesizability. Conversely, for therapeutic peptides, the focus shifts to generating functional sequences and designing de novo structures, where the primary challenges are achieving biological stability against proteolysis, ensuring proper folding, and minimizing immunogenicity. Despite these distinct challenges, both domains face shared hurdles: the scarcity of high-quality experimental data, the reliance on inaccurate scoring functions for validation, and the crucial need for experimental validation. We conclude that the full potential of diffusion models will be unlocked by bridging these modality-specific gaps and integrating them into automated, closed-loop Design-Build-Test-Learn (DBTL) platforms, thereby shifting the paradigm from mere chemical exploration to the on-demand engineering of novel therapeutics. Full article

To investigate the regulatory effects of dietary fiber structure (β-glucan-to-arabinoxylan ratio, β/AX) and apparent viscosity (AV) on production performance in pigs, this study used a 2 × 3 factorial design, randomly assigning 36 growing pigs (47.2 ± 1.5 kg) to six dietary treatments (two AV levels and three β/AX ratios), and observed the growth performance, carcass traits, meat quality, intestinal microbiota, and liver lipid metabolism. The results showed that increased dietary β/AX and AV reduced subcutaneous fat deposition, improved meat tenderness and the nutrient content of meat, but decreased pig weight gain and dressing percentage. Increased dietary β/AX and AV selectively promoted the relative abundance of butyrate-producing bacteria and the concentration of butyrate in the middle colon, thereby regulating the expression of genes related to hepatic de novo lipid synthesis and oxidation, reducing serum glucose and total cholesterol levels, and increasing plasma glucagon-like peptide-1. These findings reveal the potential mechanism by which the physicochemical properties of dietary fiber mediate lipid metabolism to reduce weight gain and provide new insights for regulating fat deposition in pigs by controlling the structural and physical properties of dietary fiber. Full article

Human Epidermal Growth Factor Receptor 2 (HER2) is a key therapeutic target in breast cancer. However, the application of existing anti-HER2 antibody drugs is limited by such issues as large molecular weight and poor stability. In this study, a series of small protein minibinders targeting HER2 domain IV were de novo designed using the RFdiffusion method. Candidate molecules were selected through a combination of ProteinMPNN and AlphaFold2 screening, and their binding capabilities were further evaluated using Escherichia coli surface display coupled with flow cytometry analysis. By integrating molecular dynamics simulations, confocal fluorescence imaging, and isothermal titration calorimetry (ITC) experiments, a highly efficient minibinder (0_703_6) with nanomolar affinity and a smaller molecular size was finally identified. Compared with the existing drug molecules, the identified minibinder exhibited approximately threefold higher affinity and a threefold reduction in molecular size. This study provides strong support for the development of novel, stable, and easily expressible HER2-targeted therapeutic molecules and also offers new insights into the rapid development of robust breast cancer drugs that may serve as ideal alternatives to monoclonal antibodies. Full article