Recep Adiyaman
bioinformatics

Issue #41: High-accuracy protein complex structure modeling based on sequence-derived structure complementarity.

February 05, 2026 Daily Intelligence
Protein Design Daily

Building something in Protein Design?

I love collaborating on new architectural challenges. Let's build together.

LinkedIn Connect Direct Message

🧬 Protein Design Digest

Curated protein signals by Recep Adiyaman

Join 1,000+ researchers. Unsubscribe anytime.

🚀 Today’s Top Signal

High-accuracy protein complex structure modeling based on sequence-derived structure complementarity.

🧬 Abstract

In living organisms, proteins perform key functions required for life activities by interacting to form complexes. Determining the protein complex structure is crucial for understanding and mastering biological functions. Although AlphaFold2 makes a revolutionary breakthrough in predicting protein monomeric structures, accurately capturing inter-chain interaction signals and modeling the structures of protein complexes remain a formidable challenge. In this work, we report DeepSCFold, a pipeline for improving protein complex structure modeling. DeepSCFold uses sequence-based deep learning models to predict protein-protein structural similarity and interaction probability, providing a foundation for identifying interaction partners and constructing deep paired multiple-sequence alignments (MSAs) for protein complex structure prediction. Benchmark results show that DeepSCFold significantly increases the accuracy of protein complex structure prediction compared with state-of-the-art methods. For multimer targets from CASP15, DeepSCFold achieves an improvement of 11.6% and 10.3% in TM-score compared to AlphaFold-Multimer and AlphaFold3, respectively. Furthermore, when applied to antibody-antigen complexes from the SAbDab database, DeepSCFold enhances the prediction success rate for antibody-antigen binding interfaces by 24.7% and 12.4% over AlphaFold-Multimer and AlphaFold3, respectively. These results demonstrate that DeepSCFold effectively captures intrinsic and conserved protein-protein interaction patterns through sequence-derived structure-aware information, rather than relying solely on sequence-level co-evolutionary signals.

Why it matters:

⭐ Additional Signals

Highly accurate protein structure prediction-based virtual docking pipeline accelerating the identification of anti-schistosomal compounds.

Schistosomiasis is a major neglected tropical disease that lacks an effective vaccine and faces increasing challenges from praziquantel resistance, underscoring the urgent need for novel therapeutics. Target-based drug discovery (TBDD) is a powerful strategy for drug development. In this study, we utilized AlphaFold to predict the structures of target proteins from Schistosoma mansoni and S. japonicum, followed by virtual molecular screening to identify potential inhibitors. Among 202 potential therapeutic targets, we identified 37 proteins with high-accuracy structural predictions suitable for molecular docking with 14,600 compounds. This screening yielded 268 candidate compounds, which were further evaluated ex vivo for activity against both adult and juvenile S. mansoni and S. japonicum. Seven compounds exhibited strong anti-schistosomal activity, with HY-B2171A (Carubicin hydrochloride, CH) emerging as the most potent. CH was predicted to target the splicing factor U2AF65, and knockdown of its coding gene Smp_019690 resulted in a phenotype similar to CH treatment. RNA sequencing revealed that both CH treatment and Smp_019690 RNA interference (RNAi) disrupted splicing events in the parasites. Further studies demonstrated that CH impairs parasite viability by inhibiting U2AF65 function in mRNA splicing regulation. By integrating RNAi-based target identification with structure-based virtual screening, alongside ex vivo phenotypic and molecular analyses of compound-treated schistosomes, our study provides a comprehensive framework for anti-schistosomal drug discovery and identifies promising candidates for further preclinical development.

Protein Structural Model Selection Informed by Comparison of Predicted Ligand Binding Poses.

Recent advances in protein structure prediction have highlighted the importance of a longstanding problem: given multiple structural models of a protein, how does one select the best model to use when predicting interactions between that protein and candidate drug molecules? Here we demonstrate the value of a previously unutilized source of information in addressing this problem. We show that given multiple ligands known to bind the protein, one can perform effective model selection by comparing the predicted binding poses of multiple ligands at each model. We introduce a method, RevBind, that exploits this information, leveraging the statistical tendency of different ligands to form similar chemical interactions with a protein’s binding pocket. RevBind can be used, for example, to select among variants of AlphaFold models, identifying those that are most useful for molecular docking. Our findings pave the way for the development of even better model selection methods that draw simultaneously on the information used by RevBind and the information used by previous methods.

From MM-PBSA to H-MMGB: Multiscale Modeling for Biomolecular Structure and Drug Discovery.

From early efforts to predict protein structure from simplified models, computational biophysics has progressed toward increasingly physics-based approaches for evaluating biomolecular structure, molecular interactions, and energetics. The molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) method provided one of the first broadly accessible ways to evaluate binding and folding energetics from molecular dynamics (MD) trajectories, with applications ranging from protein structure prediction benchmarks to protein-ligand affinity ranking. Building on this foundation, the hierarchical Molecular Mechanics Generalized Born (H-MMGB) approach was developed to provide MMGB-based binding free energy estimates more efficiently, employing the Generalized Born model in contrast to the Poisson-Boltzmann framework of MM-PBSA and thereby enabling prospective applications to ligand design. Case studies illustrate how these methods, ranging from protein folding assessment to intact-ligand modeling and to a deconstruction-reconstruction strategy using picofragments, enable hypothesis generation in the absence of experimental structures and in challenging protein-protein interaction targets. Together, these developments support a guiding principle: gradual incorporation of more physics into modeling workflows increases the probability of successfully meeting objectives across diverse computational simulation problems.

🧪 AI & Research News

🏢 Industry Insight & Applications

⚡ Quick Reads

Study on the Mechanism of Ku Diding in the Treatment of Diabetes based on Network Pharmacology, Molecular Docking Technology, and Molecular Dynamics.

Introduction To explore how Ku Diding (KDD) works in managing Diabetes Mellitus (DM), researchers utilized network pharmacology, molecular docking, and molecular dynamics methodologies. Methods Key active components of KDD were identified using the Traditional Chinese Medicine Systematic Pharmacology Database and Analysis Platform (TCMSP). Data for diabetesrelated targets were retrieved from the Human Genetic Comprehensive Databases (Genecards) and the Online Mendelian Inheritance in Man (OMIM) database. The intersection of these targets was analyzed to determine potential therapeutic targets for diabetes treatment. Proteinprotein interaction networks (PPI) were constructed using the STRING database and Cytoscape software, followed by Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. Molecular docking between the components and key targets was performed using the AutoDock Vina platform. Results This study identified that Dihydrosanguinarine, (S)-Scoulerine, among others, are the main active ingredients of KDD for treating DM, showing high affinity for critical targets like PTGS2 and PRKACA, through multiple pathways including vascular regulation, neuromodulation, metabolic regulation, and endocrine regulation. The molecular docking results showed that there are interactions between the active ingredients and the key targets, with the majority of the effective components exhibiting a stronger binding affinity than Metformin. Among them, (S)-Scoulerine and Dihydrosanguinarine demonstrated high docking affinity with the key target proteins PTGS2 and PRKACA. Discussion DM is closely linked to oxidative stress, chronic inflammation, and insulin signaling dysregulation. This study reveals that KDD exerts anti-diabetic effects via a multi-target network involving proteins such as PRKACA, PTGS2, ESR1, FOS, and DRD2. These targets are associated with glucose metabolism, inflammation, oxidative stress, and neural regulation. Modulation of these pathways likely enhances insulin sensitivity, lowers blood glucose, suppresses inflammation, and protects against oxidative damage. GO and KEGG analyses further indicate involvement in MAPK signaling, synaptic transmission, and vascular regulation, forming a multidimensional “metabolism-inflammation-neural” regulatory network. Compared to Metformin, most KDD-derived compounds showed stronger binding, highlighting their therapeutic potential. Molecular dynamics simulations support the stability of the observed binding conformations, suggesting their potential as therapeutic targets. These findings underscore KDD’s ability to simultaneously target multiple pathological mechanisms, offering a holistic treatment strategy for DM. Conclusion This study provides preliminary evidence that KDD is characterized by a multicomponent, multi-target, and multi-pathway approach in the treatment of diabetes mellitus (DM), thereby establishing a scientific foundation for further in-depth exploration of KDD’s molecular mechanisms.

High-accuracy protein complex structure modeling based on sequence-derived structure complementarity.

In living organisms, proteins perform key functions required for life activities by interacting to form complexes. Determining the protein complex structure is crucial for understanding and mastering biological functions. Although AlphaFold2 makes a revolutionary breakthrough in predicting protein monomeric structures, accurately capturing inter-chain interaction signals and modeling the structures of protein complexes remain a formidable challenge. In this work, we report DeepSCFold, a pipeline for improving protein complex structure modeling. DeepSCFold uses sequence-based deep learning models to predict protein-protein structural similarity and interaction probability, providing a foundation for identifying interaction partners and constructing deep paired multiple-sequence alignments (MSAs) for protein complex structure prediction. Benchmark results show that DeepSCFold significantly increases the accuracy of protein complex structure prediction compared with state-of-the-art methods. For multimer targets from CASP15, DeepSCFold achieves an improvement of 11.6% and 10.3% in TM-score compared to AlphaFold-Multimer and AlphaFold3, respectively. Furthermore, when applied to antibody-antigen complexes from the SAbDab database, DeepSCFold enhances the prediction success rate for antibody-antigen binding interfaces by 24.7% and 12.4% over AlphaFold-Multimer and AlphaFold3, respectively. These results demonstrate that DeepSCFold effectively captures intrinsic and conserved protein-protein interaction patterns through sequence-derived structure-aware information, rather than relying solely on sequence-level co-evolutionary signals.

Exploring toxicity and mechanisms of DDTs in Alzheimer’s disease through network toxicology and molecular docking insights.

BackgroundDichlorodiphenyltrichloroethane (DDT) and its metabolites (DDTs), such as dichlorodiphenyldichloroethylene (DDE) and dichlorodiphenyldichloroethane (DDD), are synthetic organochlorine pesticides with long environmental persistence. Although DDT has been phased out in many countries, DDE and DDD remain prevalent worldwide. Growing evidence links DDTs exposure to Alzheimer’s disease (AD), though underlying molecular targets and mechanisms remain unclear.ObjectiveIn this study, we investigated molecular targets and pathways through which DDTs potentially induce AD using network toxicology combined with molecular docking techniques.MethodsAD-related targets associated with DDTs were identified through bioinformatics searches. Key targets were selected via STRING protein-protein interaction analysis and Cytoscape, followed by signaling pathway enrichment analysis. Diagnostic efficacy was evaluated using ROC curve analysis and nomogram modeling based on GEO datasets. Molecular docking validated binding affinity between DDTs and core target proteins predicted by AlphaFold 3.ResultsWe identified 1732 potential molecular targets linking DDTs exposure to AD. Pathway analysis revealed DDTs predominantly affect AD pathogenesis by modulating apoptosis, p53 signaling, TNF signaling, and IL-17 signaling pathways. STRING and Cytoscape analyses identified seven core targets. GEO dataset validation indicated RPL23, RPS6, and RPS8 as pivotal targets, with RPL23 having strongest predictive capacity. Molecular docking confirmed binding interactions between DDTs and RPL23, with binding energies of -7.2 kcal mol -1 for DDT, -6.5 kcal mol -1 for DDE, and -7.1 kcal mol -1 for DDD.ConclusionsThis research provides novel insights into neurotoxic mechanisms of DDT and its persistent metabolites DDE and DDD, and supports enhanced public health strategies for AD prevention.

Molecular docking and dynamics in protein serine/threonine kinase drug discovery: advances, challenges, and future perspectives.

Protein serine/threonine kinases (STKs) regulate critical signaling pathways involved in cell growth, proliferation, metabolism, and apoptosis. Aberrant kinase activity is implicated in diverse human diseases, including cancer, neurodegeneration, and inflammatory disorders. Structure-based drug discovery, utilizing molecular docking and molecular dynamics (MD) simulations, has become a central strategy for identifying and optimizing STK inhibitors. In this review, we summarize recent advances and challenges in applying these in silico approaches to STK drug discovery. We discuss the principles, performance, and limitations of docking and MD approaches, as well as their integration with binding free-energy estimation methods. We emphasize recent methodological progress, including automated MD workflows, machine learning-driven interaction fingerprinting frameworks, and the growing adoption of hybrid docking-MD pipelines that enhance throughput and reproducibility. The review also highlights emerging directions such as computational design of heterobifunctional degraders (PROTACs) and allosteric modulators, which extend the scope of kinase targeting beyond ATP-competitive inhibitors. Quantitative examples of computational resource requirements and hit-validation rates from representative studies are summarized to contextualize the predictive power and practical feasibility of these approaches. Together, these developments demonstrate how the synergy of physics-based simulations, enhanced sampling, and machine learning is transforming MD from a purely descriptive technique into a scalable, quantitative component of modern kinase drug discovery.

High-resolution protein modeling through Cryo-EM and AI: current trends and future perspectives - a review.

The structural elucidation of proteins is fundamental to understanding their biological functions and advancing drug discovery. Recent breakthroughs in cryo-electron microscopy (cryo-EM) and artificial intelligence (AI)-based structure prediction have revolutionized protein modeling by enabling near-atomic resolution visualization and highly accurate computational predictions from amino acid sequences. This review summarizes the latest advances that have transformed protein structural biology from a predominantly structure-solving endeavour to a discovery-driven science. We discuss the complementary roles of cryo-EM and AI, including developments in direct electron detectors, advanced image processing, and deep learning algorithms exemplified by AlphaFold 2 and the emerging AlphaFold 3. These technologies facilitate detailed insights into challenging protein targets such as membrane proteins, flexible and intrinsically disordered proteins, and large macromolecular complexes. Furthermore, we highlight applications of integrative approaches in drug design, enzymatic mechanism elucidation, and functional predictions, illustrated by examples including hemoglobin, which demonstrates both the strengths and current limitations of AI-cryo-EM integration, and cytochrome P450 enzymes, where AlphaFold predictions have been combined with cryo-EM maps to explore conformational diversity. The review also addresses ongoing challenges and promising future directions for integrating experimental and computational methods to accelerate the exploration of protein structure-function relationships, ultimately impacting biomedical research and therapeutic development.

The promise of AlphaFold for gene structure annotation

Background As sequencing technology improves, more genomes become available. Most lack annotation, automated methods are error prone, and few genomes are ever manually curated due to time and cost. Protein structure prediction software may provide new angles for assessing and improving gene models without requiring experimental data. In this paper, we explore whether scores from protein structure prediction can aid in scoring gene model quality. We chose three species ( Fusarium graminearum , Toxoplasma gondii , and Aspergillus fumigatus ) from the VEuPathDB database which have collectively undergone more than 1000 manual curation events. We modelled translations of the gene models with AlphaFold 3, before and after curation, collecting various scores. Then we carried out structure searching of the PDB with Foldseek and sequence-based domain identification using InterProScan. We profiled the scores produced by these methods to identify those best for gene model assessment. Results AlphaFold 3 scores strongly favoured manually improved over pre-improvement models, supporting 75% of manually-curated changes in F. graminearum , 65% in T. gondii , and 84% in A. fumigatus (the lower percentage in T. gondii attributed to a high level of disorder). Further, combining scores across multiple tools (AlphaFold 3, Foldseek and InterProScan) provided additional improvements in model scoring. Conclusion Overall, the most discriminative scores combined outputs of AlphaFold 3 and Foldseek. Our results therefore highlight the potential of scores derived from deep learning-based protein structure prediction for scoring gene models in the absence of experimental data. Future work should focus on intrinsically disordered regions and developing integrated tools to apply this approach.

Assessing the Neurotoxicity of Bisphenol A Using Network Toxicology, Molecular Docking, and Molecular Dynamics Simulation.

Bisphenol A (BPA) is a prevalent environmental endocrine disruptor with potential impacts to the neurological system in humans. This study used an integrated method combining network toxicology, molecular docking, and molecular dynamics simulations to explore the molecular mechanisms underlying BPA-induced neurotoxicity. We identified 255 potential neurotoxicity-related targets through the integration and comprehensive analysis of multiple data sources, including the Comparative Toxicogenomics Database (CTD), ChEMBL, STITCH, GeneCards, and the Online Mendelian Inheritance in Man (OMIM) database. Analysis of the protein-protein interaction (PPI) network unveiled 52 core targets, among which TNF, TP53, INS, ESR1, and PTGS2 emerged as pivotal hubs in the toxicity network. Functional enrichment analysis indicated that the core targets of BPA’s influence on neurotoxicity are predominantly enriched in vital signaling cascades, including inflammatory responses, pathways of neurodegeneration, MAPK signaling pathway, serotonergic synapse pathway, and pathways in cancer. Molecular docking results demonstrated that BPA exhibited stable binding interactions with core targets. Furthermore, molecular dynamics simulations provided insights into the interactions between BPA and key targets (ESR1, TNF, and TP53), supporting the potential conformational stability of these complexes. Collectively, these computational findings contribute to understanding the potential molecular mechanisms of BPA-induced neurotoxicity and are informative for generating hypotheses related to its pathogenesis.

Network pharmacology, molecular docking, and molecular dynamics simulations to explore the effects of sinomenine on thyroid dysfunction.

Thyroid dysfunction is a disease closely associated with autoinflammatory responses and immune imbalances. Derived from the traditional Chinese medicine Sinomenium acutum, sinomenine is an alkaloid that possesses immunomodulatory and anti-inflammatory activities. However, the exact molecular mechanism underlying its therapeutic effects on thyroid dysfunction has not been clarified. This study integrates network pharmacology, molecular docking and molecular dynamics simulation techniques to explore the mechanism of sinomenine on thyroid dysfunction. Databases such as GeneCards, PharmMapper, SwissTargetPrediction, and OMIM were used to screen the targets of sinomenine and thyroid dysfunction. Subsequent GO and KEGG enrichment analyses, PPI network construction, and drug-target-pathway network analysis were conducted. Molecular docking and molecular dynamics simulations were further employed for validation. The results of GO and KEGG enrichment analyses revealed that sinomenine’s mechanism of action involves cellular responses to oxidative stress caused by inflammation, the role of nuclear transcription factors in gene expression, as well as processes of cell proliferation and apoptosis. Its targets are distributed across various pathways, suggesting a complex synergistic effect of multiple pathways in its potential mechanism; Molecular docking experiments showed that sinomenine and 14-episinomenine exhibit good binding affinity with the key targets, including TNF, STAT3, NFKB1, IL6, SRC, ESR1, and MAPK8; Molecular dynamics simulations were carried out on the two most stable binding complexes. RMSD, RMSF, Rg, and SASA curves showed that both proteins, ESR1 (PDB ID:4pxm) and MAPK8 (PDB ID:4yr8), were stabilized with sinomenine. Our study implies that via the synergistic action of multiple targets and pathways, sinomenine has the potential to impact cellular proliferation and apoptosis processes in thyroid dysfunction. The findings provide a theoretical basis for investigating the molecular mechanisms of sinomenine in treating thyroid dysfunction.

💡 Pipeline Tip

Always validate pLDDT scores before using AlphaFold models for docking.

🛠️ Resources

The protein structure is the language of life; design is its poetry. — Recep Adiyaman

BS HF DK