Recep Adiyaman
bioinformatics

Issue #21: Geometric deep learning assists protein engineering. Opportunities and Challenges.

January 13, 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

Geometric deep learning assists protein engineering. Opportunities and Challenges.

🧬 Abstract

Protein engineering is experiencing a paradigmatic transformation through the integration of geometric deep learning (GDL) into computational design workflows. While traditional approaches such as rational design and directed evolution have achieved significant progress, they remain constrained by the vastness of sequence space and the cost of experimental validation. GDL overcomes these limitations by operating on non-Euclidean domains and by capturing the spatial, topological, and physicochemical features that govern protein function. This perspective provides a comprehensive and critical overview of GDL applications in stability prediction, functional annotation, molecular interaction modeling, and de novo protein design. It consolidates methodological principles, architectural diversity, and performance trends across representative studies, emphasizing how GDL enhances interpretability and generalization in protein science. Aimed at both computational method developers and experimental protein engineers, the review bridges algorithmic concepts with practical design considerations, offering guidance on data representation, model selection, and evaluation strategies. By integrating explainable artificial intelligence and structure-based validation within a unified conceptual framework, this work highlights how GDL can serve as a foundation for transparent, interpretable, and autonomous protein design. As GDL converges with generative modeling, molecular simulation, and high-throughput experimentation, it is poised to become a cornerstone technology for next-generation protein engineering and synthetic biology.

Why it matters: Critical for improving fold accuracy and reducing structural uncertainty in de novo design.

⭐ Additional Signals

Mechanisms of cellular senescence combined with molecular docking strategies: A biomarker study of potential therapeutic targets for allergic rhinitis.

Bioinformatics and molecular docking methods were used to screen potential biomarkers of cellular senescence in allergic rhinitis (Allergic rhinitis AR), which provided a theoretical basis for revealing the mechanism of AR and exploring new therapeutic approaches. Four AR-related gene chips (GSE19187, GSE43523, GSE44037, and GSE51392) were downloaded from the gene expression database (GEO) for data pooling. Screening differential genes (DEGs) were taken to intersect with cellular senescence-related genes (SRGs) to obtain differential senescence genes (DESRGs). The differential senescence genes were subjected to Gene Ontology Database (GO) functional analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, and GSEA enrichment analysis. Protein-protein interaction (PPI) networks were constructed through the STRING database, MCODE plugin weights were analyzed to identify important gene cluster modules, and Hub genes were screened using the CytoHubba plugin topological network algorithm. Hub gene protein interactions network (GeneMANIA) was constructed by the GeneMANIA database. Predict Hub gene construct mRNA-miNA-lncRNA interactions by miRanda, miRDB, miRWalk, TargetScan, and spongeScan databases; construct Hub gene transcription factor regulatory networks by TRRUST database; analyze Hub gene-drug interactions by DGIdb database and select commonly used drugs in the clinic for molecular docking validation. A total of 264 differential genes were screened in the training set with corrected P.adj < 0.05 and |log2FC| ≥ 1.2 as the filtering condition, and a total of 866 cellular senescence genes, and 20 differential senescence genes (DESRGs) were obtained by taking the intersection of the two. A total of 19 Hub genes were obtained after PPI analysis, which were CCL2, STAT1, TLR2, IGFBP3, TLR3, KLF4, IL1RN, IRF1, SERPINB2, DPP4, MME, NQO1, SAMHD1, XAF1, PHGDH, EIF4EBP1, CTH, HSPA2, AHR The gene-protein interaction network identified 19 Hub genes associated with 21 functional proteins. 5 of the Hub gene loci were associated with 29 miRNAs and 53 lncRNAs. The transcription factor regulatory network obtained 15 transcription factors capable of regulating Hub genes. The analysis of drug-gene interactions identified 489 drugs that target hub genes. For example, in the case of budesonide, the interacting genes STAT1, TLR2, TLR3, and AHR were selected for molecular docking. Similarly, for mometasone, the interacting genes TLR2 and CTH were chosen for molecular docking. Mining AR-related Hub senescence genes by bioinformatics analysis, constructing PPI network, ceRNA network, transcription factor regulatory network, gene-drug interaction network and molecular docking validation, we screened 19 CCL2, STAT1, TLR2, IGFBP3, TLR3, KLF4, IL1RN, IRF1, SERPINB2, DPP4, MME, NQO1, SAMHD1, XAF1, PHGDH, EIF4EBP1, CTH, HSPA2, and AHR are expected to be Hub genes for potential diagnostic and therapeutic biomarkers, which will provide targets and new insights for further in-depth explorations of AR cellular senescence-related mechanisms of action and therapy.

Design, Synthesis, Molecular Docking, Structure Activity Relationship, and In Vivo Evaluation of Pyrazole-Pyrimidines for Discovering New Nonsteroidal Anti-Inflammatory Drugs.

Nonsteroidal anti-inflammatory drugs are among the most prescribed worldwide to treat pain, fever, and inflammation. However, they can cause severe adverse effects such as gastric, duodenal, hepatic, and renal injuries. Thus, the search for effective and new drugs is of high priority. Herein, the synthesis of a new series 4-((5-substituted-3-(trifluoromethyl)-1H-pyrazol-1-yl)methyl)-6-(trifluoromethyl) pyrimidin-2-substituted (pyrazole-pyrimidines) obtained through the cyclocondensation reaction of pyrazole-enaminones with amidines under mild conditions is reported. The chemical structures are confirmed by 1 H and 13 C NMR, mass spectrometry, and single-crystal X-ray analysis for compounds 4c and 4g. Molecular docking studies are conducted to identify selective cyclooxygenase-2 (COX-2) inhibitors, revealing that compounds 4d, 4j, and 4k display higher binding affinity. ADMET predictions (absorption, distribution, metabolism, excretion, and toxicity) corroborate to the docking results, suggesting favorable pharmacokinetic and toxicological properties. The in vivo antinociceptive activity is investigated in mice using the capsaicin-induced nociception model. Oral administration of compounds 4d, 4e, 4f, 4j, and 4k significantly reduces nociceptive responses, achieving effects comparable or superior to celecoxib, without altering locomotor activity. Altogether, the findings demonstrate that pyrazole-pyrimidine derivatives, especially 4d and 4k, are promising candidates for the development of selective COX-2 analgesics, combining antinociceptive efficacy with a favorable toxicological profile.

Mechanistic study of plastic monomers in gestational diabetes mellitus: A network toxicology and molecular docking approach.

Plastics are widely used in various fields such as food packaging, textile fibers, building materials, and transportation. Although the relationship between plastic additives and diseases has been reported, there is limited research on the association between plastic monomers (PM) and gestational diabetes mellitus (GDM). This study aims to investigate the link between environmental PM and GDM. By employing advanced network toxicology and molecular docking techniques, we successfully elucidated the molecular mechanisms by which PM may induce GDM. Utilizing databases such as PubChem, SEA, Super-PRED, SwissTargetPrediction, PharmMapper, Gene Cards, and OMIM, we identified potential targets associated with the disease. Further analysis using STRING and Cytoscape software helped determine the core targets most significantly related to these metabolic disorders. Additionally, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses were conducted using the David database to characterize these core targets. Finally, molecular docking with CB-Dock2 was used to validate the binding affinity of PM to these target proteins. Our findings suggest that PM may potentially induce GDM by modulating the insulin signaling pathway through STAT3, AKT1, and TP53. In summary, this work provides novel insights into the mechanisms by which environmental pollutants may trigger GDM, thereby laying a theoretical foundation for disease prevention and treatment. It offers valuable references for the safety evaluation of plastics, urging food safety regulatory agencies to strengthen oversight and encouraging the public to reduce plastic usage.

🧪 AI & Research News

🏢 Industry Insight & Applications

⚡ Quick Reads

Network Pharmacology and Molecular Docking Identify Medicarpin as a Potent CASP3 and ESR1 Binder Driving Apoptotic and Hormone-Dependent Anticancer Activity.

Ovarian cancer (OC) remains one of the most lethal gynecologic malignancies due to late diagnosis, rapid progression, and frequent chemoresistance. Despite advances in targeted therapy, durable responses are uncommon, underscoring the need for novel multitarget agents capable of modulating key oncogenic networks. Medicarpin, a natural pterocarpan phytoalexin, exhibits diverse pharmacological activities; however, its molecular mechanisms in OC are poorly defined. This study employed an integrative in silico framework combining network pharmacology, pathway enrichment, molecular docking, and survival analysis to elucidate medicarpin’s therapeutic landscape in OC. A total of 107 overlapping targets were identified, resulting in a dense protein-protein interaction network enriched in kinase-mediated and apoptotic signaling pathways. Ten hub genes were emphasized: CASP3, ESR1, mTOR, PIK3CA, CCND1, GSK3B, CDK4, PARP1, CHEK1, and ABL1. Gene Ontology and KEGG analyses demonstrated substantial enrichment in the PI3K-Akt/mTOR and prolactin signaling pathways. Docking revealed the stable binding of medicarpin to CASP3 (-6.13 kcal/mol) and ESR1 (-7.68 kcal/mol), supporting its dual regulation of hormonal and apoptotic processes. Although CASP3 and ESR1 expression alone lacked prognostic significance, their network interplay suggests synergistic relevance. Medicarpin exhibits multitarget anticancer potential in OC by modulating kinase-driven and hormone-dependent pathways, warranting further experimental validation.

Immunoinformatics-based design and evaluation of a multi-epitope vaccine against Vibrio fluvialis.

Vibrio fluvialis is an emerging foodborne pathogen causing gastroenteritis and extraintestinal infections, representing a significant public health concern due to rising antimicrobial resistance and the absence of an approved vaccine. This study aimed to design a multi-epitope subunit vaccine against V. fluvialis using immunoinformatics and a standard multi-epitope vaccine design pipeline. Two surface-exposed immunogenic membrane proteins, ATP-dependent zinc metalloprotease FtsH and lytic murein transglycosylase F, were selected as antigenic targets. Ten epitopes, including four MHC class I, four MHC class II, and two B-cell epitopes, were predicted and assembled into a 246 amino acid vaccine construct. The construct showed an antigenicity score of 0.8610. Population coverage analysis indicated that these epitopes could potentially cover 99.97% of the global population. The vaccine exhibited favorable physicochemical properties, with an instability index of 33.18 and a GRAVY score of - 0.282, suggesting stability and hydrophilicity. The tertiary structure was modeled using AlphaFold3 and docked with Toll-like receptor 2, yielding a docking score of - 270.01. Molecular dynamics simulations for 100 ns suggested stability of the vaccine-TLR2 complex. Codon optimization indicated high expression potential in Escherichia coli, with a CAI value of 0.95. Overall, the vaccine showed strong in silico immunogenic potential and requires further experimental validation through in vitro and in vivo studies.

<i>In silico</i> structural analysis of carbapenemase variants in <i>Klebsiella pneumoni</i>ae: insights for precision drug discovery against multidrug-resistant strains.

Background Carbapenemase-producing Klebsiella pneumoniae severely limits treatment options by inactivating carbapenem and other β-lactam antibiotics. To support precision drug discovery, this study investigates how structural and dynamic differences among major carbapenemase families shape their interaction with carbapenem drugs. Research design and methods The authors performed an in-silico analysis of five key Klebsiella pneumoniae carbapenemases (KPC, NDM, VIM, IMP, OXA-48) using multiple sequence alignment, homology modeling, molecular docking, and 50-ns molecular dynamics simulations. Results Sequence identity between variants was low (33.8-48.5%), indicating deep evolutionary divergence. SWISS-MODEL homology models showed high stereochemical quality, supporting reliable active-site interpretation. Docking suggested stronger binding of meropenem and imipenem to KPC and NDM, mediated by conserved catalytic residues in class A (Ser70, Lys73, Glu166) and class B (His120, His122, Asp124). MD simulations indicated more rigid, compact complexes for KPC and NDM, contrasted with higher flexibility in OXA-48. Conclusions These results reinforce that Klebsiella pneumoniae carbapenemases are not interchangeable targets and that inhibitor design must account for family-specific active-site geometries and dynamics. Integrating such structural insight with future virtual screening and experimental validation could enable variant-tailored inhibitors rather than relying on a single broad-spectrum carbapenemase blocker.

Structure Elucidation, Biological and Molecular Docking Studies of the δ‑Endotoxin Cry1Ca17 from Bacillus thuringiensis Strain BUPM14

Phytochemical Profiling, Molecular Docking, ADMET Analysis of Zingiber Officinale Peel Extract for the Biogenic Synthesis of ZnO and Fe-Doped ZnO Nanoparticles with Antibacterial and Photocatalytic Potential

Alkamides from Piper nigrum and their potential inhibitory effect on NLRP3 inflammatory activation.

Four previously undescribed alkamides (1-4) were isolated from the fruits of Piper nigrum. Their chemical structures were elucidated using HRESIMS, NMR, and optical rotation analyses. A classical NLRP3 inflammasome activation model was established by priming macrophages with LPS followed by nigericin stimulation. Compounds 1-4 exhibited markedly stronger inhibition than the reference compound piperine. Molecular docking further revealed their binding modes within the NACHT domain of NLRP3. All four derivatives displayed higher docking scores than piperine, with compound 2 showing the strongest predicted binding affinity. Molecular dynamics simulations have verified that the molecule had good binding free energy with NLRP3 and indicated the key amino acids involved in the binding. These findings enrich the structural diversity of piperine derivatives and expand the molecular scaffold available for further structure-activity relationship studies, thereby providing a solid molecular basis for the rational design and synthesis of new piperine-derived NLRP3 inhibitors.

Multispectral and Molecular Dynamics Study on the Interactions Between α-Amylase and Four Sesquiterpene Lactone Compounds.

This study compared the inhibitory effects and mechanisms of four sesquiterpene lactones derived from Asteraceae plants on α-amylase using enzymatic kinetics, multi-spectral techniques, and molecular docking methods. It was found that compounds A, B, and C inhibited α-amylase activity through a competitive mechanism, while compound D exhibited non-competitive inhibition, with compound B showing the strongest inhibitory activity. Further analysis revealed that the four sesquiterpene lactone compounds interacted with key amino acid residues of α-amylase via hydrogen bonding and hydrophobic interactions, forming stable protein-ligand complexes. Among them, compound B demonstrated the lowest binding energy, indicating the strongest binding affinity to α-amylase. By investigating the inhibitory effects of these sesquiterpene lactones with a common parent structure on α-amylase, the inhibitory mechanisms and potential pharmacophoric groups were elucidated, providing a scientific foundation for the development of antidiabetic functional foods and drugs based on sesquiterpene lactone scaffolds.

A novel CLPP variant in a Pakistani family with Perrault syndrome associated with recurrent fevers.

Perrault syndrome (PRLTS) is an autosomal recessive disease with sensorineural hearing loss and ovarian dysfunction in girls, and either a fluctuating neurological phenotype or not. PRLTS type 2 is known to be caused by pathogenic variants of the CLPP gene that encodes mitochondrial ATP-dependent protease. This paper involved clinical and genetic studies on a Pakistani family with PRLTS. Whole-exome sequencing identified a novel homozygous CLPP missense mutation (NM_006012.4: c.250 A > C; p.Ile84Leu). Its pathogenicity was assessed with the help of multiple sequence alignment, AlphaFold protein modeling, and docking with CLPX with the help of ClusPro. Auditory brainstem responses and tympanometry were in clinical assessment. The individuals were found to have a uniform phenotype of severe sensorineural hearing loss, mild intellectual disability, ataxia and frequent fever. There was one patient in whom the unilateral Eustachian tube dysfunction was hinted at by Tympanometry. At the molecular level, the identified CLPP variant involved a highly conserved residue. Structural modeling showed preserved protein architecture, whereas docking simulations revealed disrupted CLPP-CLPX interaction, suggesting a basis for impaired proteostasis. We report a novel CLPP missense variant (p.I84L) in a Pakistani family with PRLTS, expanding the mutational spectrum of CLPP. To the best of our knowledge, recurrent fever was reported in PRLTS for the first time, which expanded the PRLTS phenotype spectrum.

💡 Pipeline Tip

Employ HADDOCK for ambiguous restraints in protein-protein docking.

🛠️ Resources

Deep learning is not a magic wand, but a powerful lens for structural biology. — Recep Adiyaman

BS HF DK