📋 Browse Articles

Drug-drug interactions (DDIs) are a significant source of morbidity and adverse drug events (ADEs), particularly in situations of polypharmacy and complex medication regimens. While rules-based software integrated in electronic health records (EHRs) has demonstrated proficiency in identifying DDIs present in medication regimens, large language model (LLM) based identification requires thorough benchmarking and performance evaluation using high-quality datasets for safe use. The purpose of this study was to develop a series of performance benchmarking experiments specifically for LLM performance in identification and management of DDIs using a specifically curated clinician-annotated dataset of clinically-relevant DDIs. Show less

Artificial intelligence (AI) is being used in oncological drug development to address the high costs, low success rates, and long timelines that characterize traditional drug development pipelines. The use of machine learning (ML) and deep learning (DL) models in computer-aided drug design is constantly growing owing to their capacity to analyze large, heterogeneous datasets, their ability to capture nonlinear biological trends, and their integration of various molecular and clinical characteristics. AI applications accelerate target discovery by predicting protein structures, ranking disease-relevant genes, and assessing target drugability. AI can be used to conduct rapid searches of multiplexed chemical libraries, predict drug-target interactions, and optimize the pharmacological and physicochemical properties of drugs in virtual screening. Advanced neural network designs also aid in de novo drug design, which involves developing new molecular structures with therapeutic properties of interest. This review outlines how AI has been used for target identification, virtual screening, de novo molecular design, and, specifically, in cancer applications. It further discusses the major issues in AI-based drug development, such as data quality, model interpretation, computational constraints, and ethical and regulatory considerations, which remain essential obstacles to broader clinical translation. Show less

Proteomics has matured into a discipline capable of quantifying nearly every protein encoded by the genome, yet it remains largely blind to the true operational units of physiology: proteoforms. Each proteoform—defined by a specific sequence and post-translationally modified state—represents a unique molecular identity with distinct chemical, functional, and structural properties. This review proposes the proteoform functor: a mathematical map between the abstract proteoform state space and the realised physiological space of biological function—and ultimately complex phenotypes. Show less

A model-agnostic empirical framework is proposed to measure the uncertainty associated with protein embeddings and to assess the biological relevance of these embeddings in order to improve model reliability and performance on downstream tasks. Show less

Abstract This article presents the synthesis and detailed structural investigation of a new palladium(II) heteroligand complex trans-[PdCl(CNXyl)2(C{CN(H)Xyl}2)]Cl, containing an acyclic diaminocarbene ligand and two isocyanide ligands. X-ray diffraction and NMR spectroscopy revealed that complex is stabilized both in the crystal and in solution by a system of N–H···Cl– hydrogen bonds formed between one chloride anion and two N–H groups of the diaminocarbene ligand (N–H···Cl–···H–N), with calculated energies of 3.8–5.4 kcal/mol. A Cambridge Structural Database search identified 22 other palladium(II) acyclic diaminocarbene complexes with similar N–H···X···H–N (X = Cl–, Br–, O=C<, O=S<, etc.) hydrogen-bonded systems. The complex demonstrated significant antiproliferative activity against the triple-negative breast cancer cell line MDA-MB-231, with an IC50 value of 5.55 ± 0.45 µM, which is four times higher than that of cisplatin. Show less

Ferroptosis is a type of programmed cell death characterized by accumulation of free iron, reactive oxygen species generation and lipid peroxidation and is distinct from other types of regulated cell deaths such as apoptosis, necrosis and autophagy. Ferroptosis is distinct from other programmed cell deaths for its iron dependence and its significant role in tumor suppression. Therefore, harnessing ferroptosis may offer promising avenues for cancer therapy. In the present review, the different pathways that lead to ferroptosis, the genes and transcription factors involved in both iron and lipid metabolism, as well as the impact of small‑molecule alterations on the regulation of ferroptotic cell death, were discussed. Furthermore, the emergence of combination therapies with ferroptosis‑inducing molecules that overcome resistance to conventional chemotherapy, particularly in solid tumors, were highlighted. Show less

Combining single-cell parallel profiling of genome conformation, histone modifications, chromatin accessibility and gene expression reveals dynamics and intranuclear spatial clustering of epigenome profiles, enabling sophisticated analysis of the regulatory landscape across cell types and tissues. Show less

While various metal complexes demonstrate immunogenic cell death (ICD)-inducing properties, there is a lack of studies comparing ICD properties in structurally similar complexes with different metal centers. In this study, we synthesized four structurally similar Rh(I) and Ir(I) complexes with redox-active 1,2-bis(arylimino)acenaphthene (Ar-bian) ligands and assessed their anticancer and ICD-inducing properties. Analysis of damage-associated molecular patterns (DAMPs), ROS localization and dying cell populations highlighted the distinct roles of the metal center and the ligands. Specifically, only Rh(I) complexes induced the release of the three essential DAMPs and high levels of late apoptotic cells, while the Ir(I) complexes failed to trigger crucial “eat-me” signals. This work offers valuable insights into structure–activity relationships in metal complexes in the context of ICD. Show less

Ferroptosis is a type of programmed cell death characterized by accumulation of free iron, reactive oxygen species generation and lipid peroxidation and is distinct from other types of regulated cell deaths such as apoptosis, necrosis and autophagy. Ferroptosis is distinct from other programmed cell deaths for its iron dependence and its significant role in tumor suppression. Therefore, harnessing ferroptosis may offer promising avenues for cancer therapy. In the present review, the different pathways that lead to ferroptosis, the genes and transcription factors involved in both iron and lipid metabolism, as well as the impact of small‑molecule alterations on the regulation of ferroptotic cell death, were discussed. Furthermore, the emergence of combination therapies with ferroptosis‑inducing molecules that overcome resistance to conventional chemotherapy, particularly in solid tumors, were highlighted. Show less

Mesoionic imines (MIIs) based on a 1,2,3-triazole core have been popularized in the past ca. 5 years. In this review article we discuss the synthesis, coordination ability and the structural and spectroscopic properties of this fascinating class of electronically ambivalent compounds. Apart from this, we also discuss the utility of MIIs and their compounds in directed C–H activation reactions, and in the activation and conversion of small molecules such as alkynes and CO2. Based on the current state of the art, we touch upon possible future developments of the chemistry of these classes of molecules. Show less

Many ligands are structurally rigid and well-defined, e.g. N-heterocyclic carbenes display a fan-like structure with a defined buried volume. Here, we break this dogma by introducing more flexibility around the catalytically active center by using acyclic (diamino)carbene (ADC) ligands. The ADC ligand was constructed in a straightforward protocol on the ruthenium center via methyl isocyanide coordination and subsequent reaction with amines such as pyrrolidine. Ligand flexibility in the formed (pyrrolidine)(methylamine)carbene ruthenium complex Ru-2 was demonstrated both in solution (variable temperature NMR) and in the solid state through crystallographic identification with the protic NH site oriented either distal or proximal to the ruthenium center. In contrast to their cyclic analogues, the Ru-ADC complexes are highly active in base-free transfer hydrogenation, with turnover numbers >1000. The base-free conditions allowed for the transformation of substrates with base-sensitive groups such as esters, amides, acids, and amines, substrates that typically fail to undergo transfer hydrogenation under classical conditions. The absence of base also enabled late-stage hydrogenation of more complex substrates, and it avoids complications such as corrosion attributed to KOH and related strong bases. Show less

Hybrid phosphines with anionic hard donor functions can be used to create an adaptable ligand environment for soft late transition metals. Herein, we show that the change of coordination of a diphosphine–phosphinic acid (P3OOH) in response to acid–base interactions or hydrogen bonding results in structural transformations of a disilver complex [Ag2(P3OO)2] (1) to give solvated and protonated derivatives [Ag2(P3OOH)2]2+ (2) and [Ag3(P3OO)3H]+ (3), accompanied by the alteration of the quantum yield of the solid-state photoluminescence from 0.06 up to 0.69. The related diphosphine–phosphide oxide complexes [M2(P3O)2] (M = Ag, Au) are oxidized to phosphinate compounds 2 and non-luminescent [Au2(P3OO)2H]+ (5) in the presence of triflic acid. Alternatively, [Au2(P3O)2] readily accommodates an additional gold(I) ion to yield a trinuclear cluster [Au3(P3O)2]+ (6), which is brightly sky-blue phosphorescent in the crystalline state (Φem = 0.76). The phosphide oxide group −PO in 6 is stable towards oxidation under acidic conditions in solution but undergoes protonation that results in two orders of magnitude (>170-fold) increase of the emission intensity. Complex 6 acts as a guest in the crystalline matrix of 5 due to their structural similarity and affords solid solutions with bright luminescence at a doping content of 1–2%. Show less

Proteomics has matured into a discipline capable of quantifying nearly every protein encoded by the genome, yet it remains largely blind to the true operational units of physiology: proteoforms. Each proteoform—defined by a specific sequence and post-translationally modified state—represents a unique molecular identity with distinct chemical, functional, and structural properties. This review proposes the proteoform functor: a mathematical map between the abstract proteoform state space and the realised physiological space of biological function—and ultimately complex phenotypes. This mapping is not linear or additive. Rather, it is hierarchical, nonlinear, and context-dependent, reflecting the emergent complexity of life. Without resolving proteoforms, proteomics risks describing shadows of biology rather than its material substance. Deciphering complex phenotypes, demands a shift from bulk protein averages to revealing the precise molecular identities—proteoforms—that give rise to physiology. Show less

Biomedical research benefits from the rapid growth and diversity of experimentally detected protein–protein interactions (PPIs) by gaining important biological insights. However, increasingly dense PPI networks can be challenging to interpret and apply. The 2025 update of the Integrated Interactions Database (IID) enhances accessibility and utility through several new features. We identify and incorporate network structural components from co-purified protein sets, as well as curated and predicted complexes, enabling users to explore network organization Show less

Fluorescein isothiocyanate-conjugated Annexin V in combination with propidium iodide (PI) labelling is a widely used flow cytometric assay for quantifying apoptotic and necrotic cells in anticancer studies. However, increasing evidence suggests that double-positive cells, or the Annexin V⁺/PI⁺ fraction, may represent not only late apoptosis but also different modalities of regulated cell death, including necroptosis, pyroptosis, ferroptosis, and cuproptosis. By collating findings from preclinical studies across different cancer cells, this review highlights the need for consensus in interpreting Annexin V⁺/PI⁺ populations. In the absence of molecular and/or microscopy data, this fraction is more appropriately classified as undergoing 'late-stage cell death'. In short, establishing standardised interpretive criteria is crucial to enhance understanding, facilitate cross-study comparability, and improve the translational relevance of anticancer research. Show less

Biomedical research benefits from the rapid growth and diversity of experimentally detected protein-protein interactions (PPIs) by gaining important biological insights. However, increasingly dense PPI networks can be challenging to interpret and apply. The 2025 update of the Integrated Interactions Database (IID) enhances accessibility and utility through several new features. We identify and incorporate network structural components from co-purified protein sets, as well as curated and predicted complexes, enabling users to explore network organization beyond binary interactions. Functional, pathway, and disease associations of these components can be analyzed, enabling interactions to be grouped into higher-order structures with known or provisional biological roles. Users can now filter interactions by five detection types: pairwise, co-purification, colocalization, proximity, and other evidence. To extend the value and information of predicted interactions, we include interaction interface predictions for 53 647 PPIs, generated using the MEGADOCK docking algorithm, adding molecular detail for structural biology and variant impact studies. Finally, we map PPIs to 15 immune cell types and 12 additional normal tissues, offering tissue-specific views of interaction networks increasingly relevant in disease and immunology research. IID 2025 now includes over 1 million experimentally detected human PPIs, representing an 83% increase from the previous release, alongside expanded non-human networks. The portal remains publicly available at https://ophid.utoronto.ca/iid. Show less

This study investigates the application of machine learning techniques to predict the toxicity of tetrazole derivatives, aiding in the identification of environmental risks from chemical exposure. Utilizing LD50 data sourced from the scientific literature and the ChemIDplus database, regression models were developed to forecast acute intraperitoneal toxicity in mice. A machine learning regression model for acute intraperitoneal toxicity in mice was constructed and validated on a test dataset, achieving high accuracy (R2 = 0.76 and MSE below 10−4) and surpassing most of the comparable literature models. Molecular descriptors were computed via Mordred software to explore quantitative structure–activity relationships, and additionally, the model's robustness was demonstrated by measuring the acute toxicity of tetrazole derivatives synthesized through the azido-Ugi reaction. Show less

Zinc is a crucial element in cellular processes, and its homeostasis has intricate relationships with the initiation, progression, and therapeutic intervention of cancer. Activation of the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway has been proven to be an effective strategy for cancer immunotherapy. Herein, we report four phosphorescent iridium complexes (Ir1–Ir4) with zinc chelating ligands. Among them, Ir1 can bind and image mitochondrial chelatable zinc ions via phosphorescence-lifetime responses, consequently modulating the expression of zinc-regulatory proteins. Furthermore, the in situ formed heteronuclear metal complex Ir1-Zn2 shows nuclease mimetic activities, capable of hydrolyzing mitochondrial DNA (mtDNA) to release mtDNA fragments for the activation of the cGAS-STING pathway. In conclusion, we designed a mitochondria-targeting phosphorescent Ir(III) complex with dual functions in dysregulation of zinc homeostasis and generation of nuclease in situ, which provides an innovative approach to stimulate the cGAS-STING pathway. Show less

Background: Adverse drug reactions (ADRs) are harmful side effects of medications. Social media provides real-time, patient-generated data, though its unstructured format presents challenges. Natural language processing and transfer learning offer promising solutions. Objective: This study aimed to evaluate whether transformer-based models fine-tuned on a general ADR dataset can effectively classify ADRs from tweets related to glucagon-like peptide-1 (GLP-1) receptor agonists and to benchmark their performance against state-ofthe-art large language models (LLMs). Show less

Computational drug discovery is essential for screening potential treatments and reducing the costs and time associated with proposing or combining drugs for disease management. Despite the extensive research conducted in this field, it remains an emerging area, particularly with the advent of machine learning, deep learning, and large language models (LLMs). This systematic review examines the integration of machine learning and deep learning techniques in drug discovery, concentrating on three critical areas: drug-drug interactions (DDIs), drug-target interactions (DTIs), and adverse drug reactions (ADRs). The review analyzes over 100 papers published between 2020 and 2025, categorizing the methods into deep learning, machine learning, graph learning, and hybrid models. It highlights the transformative impact of natural language processing (NLP) and LLMs in extracting meaningful insights from biomedical literature and chemical data. Furthermore, this work introduces key databases and data sets widely utilized in drug discovery. Additionally, this review identifies gaps in the existing research, such as the lack of comprehensive studies that simultaneously address DDI, DTI, and ADR extraction, and it proposes a more holistic approach to fill these gaps. The paper concludes by thoroughly evaluating various models, underscoring their performance metrics. Show less

Motivation: Proteins are of great significance in living organisms. However, understanding their functions encounters numerous challenges, such as insufficient integration of multimodal information, a large number of training parameters, limited flexibility of classification-based methods, and the lack of systematic evaluation metrics for protein question answering systems. To tackle these issues, we propose the Prot2Chat framework. Results: We modified ProteinMPNN to encode protein sequence and structural information in a unified way. We used a large language model Show less

Introduction Mitochondria are essential organelles for many aspects of cellular homeostasis. They play an indispensable role in the development and progression of diseases, particularly cancer which is a major cause of death worldwide. We analyzed the scientific research output on mitochondria and cancer via PubMed and Web of Science over the period 1990–2023. Methods Bibliometric analysis was performed by extracting data linking mitochondria to cancer pathogenesis over the period 1990–2023 from the PubMed database which has a precise and specific search engine. Only articles and reviews were considered. Since PubMed does not support analyses by countries or institutions, we utilized InCites, an analytical tool developed and marketed by Clarivate Analytics. We also used the VOSviewer software developed by the Centre for Science and Technology Studies (Bibliometric Department of Leiden University, Leiden, Netherlands), which enables us to graphically represent links between countries, authors or keywords in cluster form. Finally, we used iCite, a tool developed by the NIH (USA) to access a dashboard of bibliometrics for papers associated with a portfolio. This module can therefore be used to measure whether the research carried out is still basic, translational or clinical. Results In total, 169,555 publications were identified in PubMed relating to ‘mitochondria’, of which 34,949 (20.61%) concerned ‘mitochondria’ and ‘dysfunction’ and 22,406 (13.21%) regarded ‘mitochondria’ and ‘cancer’. Hence, not all mitochondrial dysfunctions may lead to cancer or enhance its progression. Qualitatively, the disciplines of journals were classified into 166 categories among which cancer specialty accounts for only 4.7% of publications. Quantitatively, our analysis showed that cancer/neoplasms in the liver (2569 articles) were placed in the first position. USA occupied the first position among countries contributing the highest number of publications (5695 articles), whereas Egypt came in the thirty-eight position with 84 publications (0.46%). Importantly, USA is the first-ranked country having both the top 1% and 10% impact indicators with 207 and 1459 articles, respectively. By crossing the query ‘liver neoplasms’ (155,678) with the query ‘mitochondria’ (169,555), we identified 1336 articles in PubMed over the study period. Among these publications, research areas were classified into 65 categories with the highest percentage of documents included in biochemistry and molecular biology (28.92%), followed by oncology (23.31%). Conclusions This study underscores the crucial yet underrepresented role of mitochondria in cancer research. Despite their significance in cancer pathogenesis, the proportion of related publications remains relatively low. Our findings Supplementary Information The online version contains supplementary material available at https://​doi.​org/​10.​1007/​s12672-​025-​ 02139-5. * Abeer El Wakil, abeer_elwakil@alexu.edu.eg; Patrick Devos, patrick.devos@univ-lille.fr; Heba Abdelmegeed, hn.abdelmegeed@ nrc.sci.eg; Alaa Kamel, alaa.kamel_pg@alexu.edu.eg | 1Department of Biological and Geological Sciences, Faculty of Education, Alexandria University, Alexandria 21526, Egypt. 2Université Lille, Lillometrics, 59000 Lille, France. 3CHU Lille, Direction de la Recherche et de l’Innovation, 59000 Lille, France. 4Department of Chemistry of Natural Compounds, National Research Centre, Giza, Egypt. 5Department of Zoology, Faulty of Science, Alexandria University, Alexandria, Egypt. Discover Oncology (2025) 16:517 | https://doi.org/10.1007/s12672-025-02139-5 Vol.:(0123456789) Research Discover Oncology (2025) 16:517 | https://doi.org/10.1007/s12672-025-02139-5 highlight the need for further research to deepen our understanding of mitochondrial mechanisms in cancer, which could pave the way for new therapeutic strategies. Graphical Abstract Show less

The clinical success of metal-based anticancer agents can be achieved by developing not only an efficient metallodrug but also a suitable drug delivery system (DDS). Although spatiotemporal delivery, enhancing the efficacy, and alleviating toxicity are achievable, modifying the mechanism of action of metallodrugs using a nano DDS remains scarce. With all this in mind, a series of cyclometalated ruthenium(II) half-sandwich complexes of the type [(η⁶-p-cymene)Ru(L)Cl] Ru(1)-Ru(4), where L is 2-phenylquinoline (L1), 2-(thiophen-2-yl)quinoline (L2), 4-methyl-2-phenylquinoline (L3), or 2,4-diphenylquinoline (L4), have been isolated and characterized by analytical and spectroscopic methods. Ru(1) and Ru(2) have been structurally characterized, and their coordination geometries around the ruthenium(II) are described as pseudo-octahedral geometry. Only the Ru(1) complex, which exhibited substantial cytotoxicity in non-cancerous cells and low cytotoxicity in breast cancer cells, is encapsulated into a hybrid nanosystem comprising phospholipid and polydiacetylene. The Ru(1)-entrapped nanoassembly (PDL-Ru(1)) is found to show pH-induced emission and higher release of the complex in a simulated tumor environment than in a physiological environment. Even though such a halochromic character failed to benefit cell imaging, the nanocarrier-mediated delivery has been proven to improve the cytotoxicity of Ru(1) in breast cancer cells, modulate the mode of cell death, and reduce toxicity in normal cells. Zebrafish embryo toxicity studies revealed that polydiacetylene-lipid nanoassembly could be useful for in vivo biocompatibility applications of ruthenodrug candidates. Show less

The diversification of ligands provides more opportunities to adjust the photophysical performance as well as the bio-function of Ru(II) complexes as novel photosensitizers. Herein, a kind of Ru(II) complexes carrying resveratrol derivative, amino-Res, as ligand was designed and synthesized. The representative complex (named Ru4) showed potent anticancer activity under the trigger of 520 nm-light. Lipophilicity and cellular accumulation experiments indicated that Ru4 possessed higher LogP_O/W value and cell up-take than Ru1-Ru3 and [Ru(bpy)₃]²⁺. Mechanism study revealed that Ru4 could inhibit cancer cell migration, invasion and cancer stemness. The bio-function of Ru4 was mainly inherited from the amino-Res ligand. The in vivo study demonstrated that Ru4 could inhibit the tumor growth without significant system toxicity. Show less

Claudin (CLDN) proteins are extensively studied due to their critical role in maintaining tissue barriers and cell polarity. However, significant gaps remain in understanding the functional mechanisms of their sequence motifs and the molecular mechanisms of their interactions with other tight junction proteins. This review systematically examines the multifunctional properties of the CLDN protein family from the perspectives of sequence and structure. During evolution, CLDN family members have developed highly conserved structural features, particularly key conserved sites within the first extracellular loop (ECL1) and the C-terminal PDZ-binding domain, which play a central role in regulating the barrier function of tight junctions, ion selectivity, and protein-protein interactions. Furthermore, the distribution pattern of acidic and basic amino acids in ECL1 has been shown to directly determine ion selectivity and paracellular permeability. Meanwhile, the assembly and functional stability of tight junctions are precisely regulated by the C-terminal PDZ-binding domain through its interactions with the ZO protein family. Additionally, the study further elucidates how CLDN proteins modulate critical signaling pathways governing cellular proliferation, survival, and permeability, thereby participating in diverse physiological and pathological processes. These insights have deepened the understanding of the functional diversity of CLDN proteins and provided a new theoretical basis for developing disease diagnostic markers and designing targeted treatment strategies based on CLDN proteins. Show less