A new study published in the journal Molecular Diversity by Rujia Zhao, Tianzhu Shen, and Tongzhou Huang tackles this challenge with an ambitious computational pipeline that fuses machine learning, molecular docking, molecular dynamics simulations, and pharmacokinetic prediction into a single end-to-end workflow. Rather than relying on any one technique, the researchers built a layered funnel that begins with a curated dataset of 3,660 known SERT inhibitors and ends with three structurally distinct lead molecules whose predicted binding to the transporter is, on paper, stronger than that of escitalopram, one of the most widely prescribed antidepressants in the world. The work is entirely computational, but it offers a rigorous demonstration of how artificial intelligence can compress years of trial-and-error chemistry into a matter of weeks.

The foundation of the study is data curation. The team harvested bioactivity data from the ChEMBL database, targeting the human serotonin transporter under the identifier CHEMBL228, and applied strict filters: only entries reporting IC50 values with precise standard relations and nanomolar units were retained. After removing redundant structures and those lacking usable chemical notation, 3,045 unique molecules remained. To broaden the chemical diversity of the training set, the researchers manually extracted an additional 615 data points from seven representative patent documents, rendering the molecular structures in ChemDraw and cross-referencing them with reported potencies. The final dataset of 3,660 compounds was converted to pIC50 values, the negative decadic logarithm of IC50, to ensure a distribution suitable for regression modeling.

Representing molecules for machine learning required a dual strategy. Each compound was encoded with Morgan circular fingerprints, with a radius of 3 and a bit length of 1,024, which capture local atomic neighborhoods by iteratively traversing bonds and hashing the resulting substructures into fixed-length bit vectors. These fingerprints were concatenated with thirteen physicochemical descriptors, including molecular weight, lipophilicity, hydrogen bond donor and acceptor counts, topological polar surface area, rotatable bond counts, and the quantitative estimate of drug-likeness known as QED. The resulting feature vectors bridge structural topology and pharmacological behavior, giving the algorithms both the fine detail of chemical motifs and the coarse properties that govern absorption and binding.

Twelve regression algorithms were benchmarked across four families: linear and regularized models, kernel-based and instance-based methods, tree-based ensembles, and a multilayer perceptron neural network. The clear winners were the gradient-boosting and bagging ensembles. LightGBM achieved the best performance with an R-squared of 0.7138 and a mean squared error of 0.3969 on the held-out test set, followed closely by Random Forest at 0.7001 and XGBoost at 0.6920. Linear models fared poorly, with ordinary least squares regression producing particularly large prediction errors, a result the authors interpret as evidence that the structure-activity relationship of SERT inhibitors is fundamentally nonlinear. Importantly, the researchers are candid about the limits of these numbers: an R-squared near 0.7 is adequate for ranking and enriching compound libraries but not for quantitatively predicting exact potency, and the random data split used may overestimate generalization to entirely novel chemotypes. The models were therefore treated as prioritization engines rather than oracle-like activity predictors.

With three validated models in hand, the team unleashed them on a staggering library of 11,526,814 commercially available compounds drawn from eight chemical databases. A conservative consensus rule was applied: only compounds predicted by all three models to have a pIC50 greater than 7, corresponding to an estimated IC50 below 100 nanomolar, were allowed through. This filter reduced the library to 10,647 candidates. The authors acknowledge the trade-off inherent in such stringency, since true actives correctly flagged by only one or two models would be discarded as false negatives, but the consensus approach suppresses model-specific errors and enriches for genuinely high-affinity chemotypes. An applicability-domain analysis using Tanimoto similarity to the training molecules confirmed that the screening output remained chemically anchored: under a moderately inclusive threshold, more than 80 percent of the consensus candidates showed at least moderate similarity to known training compounds, while several top-ranked leads displayed lower similarity, consistent with desirable scaffold novelty.

The next stage was structure-based. The 10,647 candidates were prepared in three dimensions, with tautomers and protonation states generated at physiological pH, and docked into the crystal structure of human SERT, PDB entry 5I71, using a hierarchical Glide protocol that progressed from high-throughput virtual screening through standard precision to extra precision. This cascade winnowed the pool to 288 compounds, which were then refined with the Prime MM/GBSA method to estimate binding free energies. A cutoff of minus 55 kilocalories per mole yielded 58 candidates, and in silico ADMET profiling of those compounds ultimately produced a shortlist of 24 hits with generally acceptable predicted pharmacokinetic profiles. As a sanity check, the models predicted pIC50 values for escitalopram between 7.585 and 7.739, comfortably within the experimentally reported range of 2.1 to 55 nanomolar for that drug, lending credibility to the entire framework.

Five compounds emerged as the most promising: PB49039490 from UkrOrgSynthesis, compound 19,835,875 from ChemBridge, and Z105894806, Z2215663922, and Z310319934 from Enamine. Docking analysis revealed that all five occupy the same orthosteric central S1 pocket of SERT and share a conserved interaction network despite their diverse scaffolds. Each ligand forms ionic or hydrogen-bond contacts with Asp98, a residue long recognized as critical for antidepressant recognition, and engages in pi-pi stacking or cation-pi interactions with aromatic residues including Tyr95, Tyr176, Phe335, and Phe341. Quantitative similarity analysis against a reference panel of 1,439 experimentally validated high-affinity SERT inhibitors showed that 22 of the 24 hits had nearest-neighbor Tanimoto coefficients below 0.40, and similarity to escitalopram itself was strikingly low, with a median of just 0.089. In other words, these are genuinely novel scaffolds, not minor variations on existing drugs.

To test whether these binding poses would survive the thermal noise of a real molecular environment, the researchers ran molecular dynamics simulations in AMBER 22, solvating each protein-ligand complex in a TIP3P water box and performing three independent 100-nanosecond production runs per complex, for a total of 300 nanoseconds of sampling each. The results were reassuring. Protein backbone RMSD values equilibrated at approximately 2.4 to 3.3 angstroms over the final 50 nanoseconds without progressive drift, ligand RMSD values settled into a narrow window of roughly 0.8 to 1.3 angstroms, and the radius of gyration remained stable at about 24.0 to 24.5 angstroms, indicating no structural expansion or destabilization of the transporter. MM/GBSA binding free energies calculated from the trajectories placed Z2215663922 at minus 51.69 kilocalories per mole, compound 19,835,875 at minus 52.17, and Z310319934 at minus 44.48, all more favorable than the minus 40.25 kilocalories per mole computed for escitalopram. Per-residue energy decomposition highlighted Tyr95, Ile172, and Tyr176 as the dominant energetic contributors, with Tyr95 and Ile172 each contributing more than 4 kilocalories per mole in the Z2215663922 complex.

The study is honest about its limitations: the simulations used a single crystal structure without an explicit lipid membrane, entropy terms were omitted from the free-energy calculations, and no laboratory experiments were performed. ADMET predictions also flagged potential hERG cardiac liabilities for two candidates, Z2215663922 and Z105894806, whose predicted potassium channel inhibition values approach the conventional concern threshold. Still, the work delivers exactly what early-stage discovery needs: a transparent, multi-layered prioritization framework and three structurally diverse lead molecules that now merit in vitro validation in serotonin uptake assays, followed by selectivity profiling against the related noradrenaline and dopamine transporters. If even one of these computational leads survives the transition from silicon to bench, it would stand as a compelling case study for machine learning as a genuine engine of antidepressant drug discovery.

Subject of Research: Machine learning-based virtual screening for novel serotonin transporter inhibitors as potential antidepressant drug candidates

Article Title: Identification of novel serotonin transporter (SERT) inhibitors via large-scale machine learning-based virtual screening and molecular dynamics simulations

Article References: Identification of novel serotonin transporter (SERT) inhibitors via large-scale machine learning-based virtual screening and molecular dynamics simulations. (n.d.). https://doi.org/10.1007/s11030-026-11705-0

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11705-0

Keywords: serotonin transporter, SERT inhibitors, machine learning, virtual screening, molecular dynamics, drug discovery, depression, LightGBM, XGBoost, molecular docking, MM/GBSA, ADMET

The core of Meta-iPPAR is a stacking strategy, an ensemble technique in which several different machine learning algorithms are trained on the same data and a higher-level meta-model then learns how best to weigh their individual predictions. Rather than relying on a single classifier, the framework draws on a diverse set of molecular descriptors computed directly from SMILES strings, the compact text notation that encodes chemical structures. This multi-view representation allows the model to capture different aspects of a molecule’s character, from its topology and physicochemical properties to patterns of atoms and bonds. The authors, led by Phasit Charoenkwan of Chiang Mai University together with colleagues at Mahidol University, have used similar stacking approaches in previous work on antiviral peptides, ACE inhibitory peptides, and other bioactive molecules, and the new study extends that recipe into the small-molecule drug discovery arena.

The performance figures reported for Meta-iPPAR are striking. On an independent test set of compounds the model achieved an accuracy of 0.926, an area under the receiver operating characteristic curve of 0.965, and a Matthews correlation coefficient of 0.848. In practical terms, the model correctly classified more than nine out of ten compounds and showed a strong balance between sensitivity and specificity, with the MCC value indicating that its predictions were far better than chance even on a balanced problem. Metrics of this kind matter because drug discovery datasets are often skewed, and a model that simply predicts the majority class can look impressive while being useless. The combination of high AUC and high MCC suggests that Meta-iPPAR has genuine discriminative power rather than an inflated score driven by class imbalance.

What sets this work apart from many black-box machine learning studies is the emphasis on interpretability. The researchers applied SHAP analysis, a technique borrowed from game theory that assigns each molecular feature a contribution value for every prediction, allowing chemists to see which structural elements push a compound toward or away from PPAR-γ activation. Alongside this, the team carried out a scaffold analysis, breaking the active compounds down into their core molecular frameworks in the tradition of the Bemis-Murcko decomposition to identify chemotypes that recur among agonists. Together, these analyses revealed molecular features and structural motifs associated with receptor activation, giving medicinal chemists actionable design rules rather than an opaque probability score. This kind of transparency is increasingly demanded in computational drug discovery, where regulators and researchers alike want to understand why a model makes its calls.

To demonstrate the framework’s real-world utility, the team turned Meta-iPPAR loose on a large library of natural products. More than 36,000 compounds from the Natural Products Atlas, a curated database of microbially derived molecules, were screened through the model. From this virtual haystack, three fungal-derived candidates emerged as the most promising hits. The choice of a natural products library is significant: compounds made by fungi and other microbes have evolved to interact with biological targets and have historically been an extraordinarily rich source of drugs, yet they remain underexplored relative to synthetic libraries because of the difficulty of sourcing and characterizing them.

The three candidates did not rest on the machine learning prediction alone. The researchers subjected them to molecular docking, computationally fitting each molecule into the ligand-binding domain of PPAR-γ to see how and where it would sit within the receptor’s pocket. They then ran molecular dynamics simulations lasting 300 nanoseconds for each candidate, watching how the protein-ligand complexes behaved over time in a simulated watery environment. Stable binding conformations persisted throughout the simulations, and the ligands maintained favorable interactions with key residues in the ligand-binding domain. Crucially, the behavior of the predicted agonists was comparable to that of known PPAR-γ agonists and co-crystal ligands whose binding modes have been determined experimentally, lending credibility to the computational hits.

The biology behind the target explains why so much effort is being invested. PPAR-γ is the receptor engaged by the thiazolidinedione class of antidiabetic drugs such as pioglitazone, which improve insulin sensitivity but carry side effects including weight gain and cardiovascular concerns. Structural studies have shown that full agonists stabilize one conformation of the receptor’s activation helix, while partial agonists achieve beneficial effects through alternative binding modes, potentially offering a better therapeutic profile. Recent research has also focused on ligands that block the phosphorylation of a specific serine residue on the receptor, a mechanism linked to insulin resistance without the classical agonist side effects. A reliable computational filter for agonist activity could therefore accelerate the search for next-generation PPAR-γ modulators with improved safety margins.

Meta-iPPAR also arrives amid a broader wave of machine learning tools aimed at this receptor. Other groups have recently described predictors such as PPGBioPred and PGMP_v1, and integrated virtual screening campaigns combining fragment molecular orbital calculations, docking, and dynamics have been used to find partial agonists for type 2 diabetes. The Thai team’s contribution is the stacking architecture paired with interpretability and a full structure-based validation chain, a combination they argue makes the framework particularly suitable for the early stages of a drug development pipeline. The model and all underlying data have been made publicly accessible through a web interface, lowering the barrier for other laboratories to apply or benchmark it.

The authors are careful to note the limits of what has been achieved. The three fungal candidates remain computational predictions, and the study’s conclusions call for future experimental validation and biological evaluation to confirm whether the molecules truly activate PPAR-γ in cells and animals. Even so, the work illustrates a template that is rapidly becoming standard in early drug discovery: a fast, interpretable machine learning model to triage enormous chemical libraries, followed by docking and dynamics to interrogate the survivors at atomic resolution. If the fungal hits survive laboratory testing, Meta-iPPAR will have demonstrated not just a clever algorithm but a shortcut from database to drug lead, one that could compress years of trial-and-error screening into weeks of computation.

Subject of Research: Interpretable machine learning and molecular simulation for predicting PPAR-γ agonists

Article Title: Interpretable QSAR modelling for PPAR-γ agonist prediction by integrating a stacking strategy, docking, and MD simulations

Article References: Charoenkwan, P., Meewan, I., Schaduangrat, N., & Shoombuatong, W. (2026). Interpretable QSAR modelling for PPAR-γ agonist prediction by integrating a stacking strategy, docking, and MD simulations. Molecular Diversity. https://doi.org/10.1007/s11030-026-11700-5

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11700-5

Keywords: PPAR-gamma, machine learning, QSAR, stacking ensemble, virtual screening, molecular docking, molecular dynamics, SHAP interpretability, natural products, drug discovery, type 2 diabetes, fungal metabolites

The study, published in the journal Molecular Diversity, was led by Binglin Huang and Bijuan Lin of Fujian Medical University Union Hospital, together with colleagues at Fujian Medical University and Nanjing University of Chinese Medicine. Rather than relying on the well-characterized binding groove that VHL uses to recognize hydroxylated HIF-1α, the team took a different route. They used DDMut-PPI, a graph-based deep-learning tool that predicts how mutations at a protein-protein interface affect the interaction, to nominate residues that might serve as auxiliary contact points beyond the canonical binding site. Those predicted residues were then used to build an alternative pharmacophore model, a simplified three-dimensional map of the chemical features a molecule would need to wedge itself into the interface and interfere with the HIF-1α/VHL handshake.

With that model in hand, the researchers screened a library of 24,893 molecules computationally, using docking tools including AutoDock4, which allows selective receptor flexibility during the docking calculation. Candidates that docked favorably against the alternative pharmacophore were then carried into the laboratory, where a fluorescence-polarization assay measured how well each compound competed with a labeled HIF-1α peptide for binding to VHL. Four molecules emerged with half-maximal inhibitory concentrations, or IC50 values, below 10 micromolar, a respectable starting point for compounds discovered without any structure-guided medicinal chemistry.

The standout was a compound the authors labeled Cmpd16, which turned out to be ixazomib, a proteasome inhibitor already approved for use in multiple myeloma under the trade name Ninlaro. In the fluorescence-polarization assay, Cmpd16 showed the highest affinity of the four hits, with an IC50 of 0.41 micromolar. The finding that an existing drug can bind VHL and interfere with HIF-1α recognition is notable because it opens the possibility of repurposing: a molecule with known pharmacokinetics, safety data, and manufacturing routes could, in principle, be repositioned for a completely different indication, provided the new activity holds up in more demanding biological systems.

The team then asked whether the compound actually did what a VHL inhibitor should do inside cells. In their reported assays, Cmpd16 showed no detectable loss of cell viability, an important early safety signal, and it produced a VHL-dependent pattern of stabilization affecting both HIF-1α and the hydroxylated form of the protein. In other words, when VHL’s grip is loosened, even hydroxylated HIF-1α, which would normally be destroyed regardless of oxygen levels, lingers in the cell. That is precisely the biochemical signature expected from a molecule that blocks the VHL recognition step downstream of the oxygen-sensing hydroxylases, and it mirrors the mechanism of dedicated VHL inhibitors developed over the past decade by academic and industrial groups.

To test whether this molecular effect translated into protection under stress, the researchers used an oxygen-glucose deprivation and reoxygenation model, a standard laboratory mimic of ischemia and reperfusion injury in which cells are deprived of both oxygen and nutrients and then resupplied. At a concentration of 10 micromolar, Cmpd16 improved the migration of endothelial cells and their ability to form tube-like structures, processes central to the formation of new blood vessels that resupply damaged tissue. The treated cells also showed increased levels of vascular endothelial growth factor, VEGF, and the glucose transporter GLUT1, both of which are canonical HIF-1α target genes that support angiogenesis and metabolic adaptation. At the same time, the compound was associated with reduced accumulation of reactive oxygen species and reduced levels of cleaved caspase-3, a hallmark of apoptosis, suggesting that the cells were not only adapting but genuinely surviving the insult better.

Because the initial hit came from a computational model rather than from a crystal structure of a ligand bound in the interface, the authors took care to interrogate how plausible their predicted binding mode really was. They ran three independently initialized molecular dynamics simulations of 200 nanoseconds each using the Desmond engine. Across the replicas, the simulations showed recurring contacts between the ligand and two residues, Pro99 and His110, in VHL, but the overall dynamics of both the protein and the ligand varied from replica to replica. The authors are explicit about what this means: the simulations support a computationally plausible orientation for Cmpd16 rather than a unique, well-defined one. That honesty matters, because molecular dynamics can be seductive, and overstating a single binding pose from a handful of trajectories is a well-known pitfall in computational drug discovery.

The team also compared two selected ten-compound panels to probe whether their deep-learning-informed pharmacophore enriched for active molecules more effectively than a conventional approach. The comparison, analyzed with a two-sided Fisher’s exact test, yielded a p-value of 0.0867, which falls short of conventional statistical significance. The authors describe this analysis as exploratory, and they further caution that the causal role of the predicted interface residues remains unproven until experimental mutagenesis is performed. These caveats are worth emphasizing: the study demonstrates a viable discovery pipeline and a promising hit, but the mechanistic details of where and how the compound binds VHL are still hypotheses awaiting direct experimental confirmation.

The broader context makes the work interesting on several fronts. The HIF pathway has long been a drug-discovery target, though most clinical success has come from a different angle: inhibitors of the prolyl hydroxylase enzymes, such as daprodustat and enarodustat, which are approved or in late-stage development for anemia in chronic kidney disease. Those drugs prevent HIF-1α from being tagged in the first place. Directly blocking the VHL-HIF interaction is a more recent strategy, and potent chemical probes with nanomolar affinities have been reported, largely built around a conserved hydroxyproline mimetic scaffold. The structural diversity of known VHL ligands has remained narrow, which is exactly the gap the new study set out to address by looking for chemically distinct molecules through an alternative interface model. Finding an approved proteasome inhibitor within that chemically distinct set is an unexpected twist, since ixazomib was designed to inhibit the proteasome’s catalytic activity, not to bind an E3 ubiquitin ligase.

There is also a pleasing irony in the biology. VHL ligands were originally developed largely as tools for targeted protein degradation, forming the binding arm of PROTAC molecules that recruit VHL to destroy disease-causing proteins. Here, the same interaction is being blocked for the opposite purpose: to prevent VHL from destroying a protein the body needs during oxygen crisis. Whether Cmpd16, or molecules inspired by it, can eventually protect tissues in living models of ischemia remains to be seen, and the distance from an endothelial-cell dish to a patient’s heart or kidney is considerable. But the study offers a concrete example of how deep-learning predictions of interface residues can reshape the search space in virtual screening, and it adds a repurposing candidate, and a set of testable residue hypotheses, to a pathway whose pharmacological importance continues to grow. The next steps, mutagenesis of the predicted contact residues, co-crystallography or cryo-EM of the ligand bound to VHL, and validation in animal models of ischemic injury, will determine whether this computational detour leads to a genuinely new way of protecting cells when oxygen runs out.

Subject of Research: Computational discovery of small-molecule inhibitors of the HIF-1α/VHL protein-protein interaction for protection of cells under hypoxic stress

Article Title: Computational discovery of HIF-1α/VHL protein–protein interaction inhibitors for hypoxic cell protection

Article References: Computational discovery of HIF-1α/VHL protein–protein interaction inhibitors for hypoxic cell protection. (n.d.). https://doi.org/10.1007/s11030-026-11706-z

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11706-z

Keywords: HIF-1α, VHL, protein-protein interaction, virtual screening, deep learning, ixazomib, hypoxia, ischemia, molecular dynamics, drug repurposing, endothelial cells, pharmacophore

The new work builds on a wave of geometric learning methods that have transformed computational docking in recent years. Classical tools such as AutoDock, AutoDock Vina, GOLD, and MedusaDock combine a conformational search over the ligand’s translational, rotational, and torsional degrees of freedom with scoring functions that approximate the energetics of binding. Machine learning scoring functions like NNScore and RF-Score later learned nonlinear relationships between structural features and binding properties, and deep models such as AtomNet and GNINA learned interaction representations directly from three-dimensional structures. More recently, methods like EquiBind, TANKBind, E3Bind, FABind, and DiffDock have moved beyond scoring toward direct prediction or refinement of ligand geometry, with DiffDock famously reformulating pose prediction as a diffusion process over translational, rotational, and torsional variables. Systems such as DynamicBind, NeuralPLexer, AlphaFold 3, FABFlex, and FlowDock have extended these ideas to settings involving protein flexibility and joint modeling of entire protein-ligand complexes.

Yet the authors of the new study identified a persistent gap in how refinement methods handle coordinate errors. A docked ligand may need a common shift in position and orientation, which is a rigid transformation applied to the whole molecule, while at the same time individual atoms may require their own local adjustments to match the true bound conformation. A global transformation alone cannot represent atom-specific corrections, and an atom-wise displacement parameterization, the approach used by the earlier graph-based method MedusaGraph, never explicitly parameterizes a shared rotation and translation for the complete ligand. Pose accuracy must also be considered alongside molecular geometry: a low coordinate error does not by itself guarantee that the ligand’s internal structure is preserved or that it avoids physically implausible overlap with the protein. These observations motivated a design that combines complementary coordinate predictions with pose supervision and additional geometric regularization.

KTransPose answers that challenge with a dual-branch architecture built on a graph neural network encoder. The protein-ligand complex is represented as a graph whose nodes are the ligand atoms and the selected protein-pocket atoms, with edges encoding molecular connectivity and interatomic distances. The encoder, which uses a hidden dimension of 256 and residual connections with GELU activations and dropout, produces a shared representation of the graph. From this representation, a global pose transformation branch pools the ligand and protein node features separately, concatenates them, and passes them through a small multilayer perceptron that outputs an axis-angle rotation vector and a translation vector. The rotation is converted to a valid rotation matrix using Rodrigues’ formula, and the transformation is applied around the ligand centroid. Meanwhile, a local coordinate correction branch applies an additional graph layer with three output channels, producing an independent three-dimensional adjustment for every ligand atom. The final displacement for each atom is simply the sum of the global and local contributions.

The training objective is where the framework earns its docking-oriented character. Rather than supervising displacements directly, the loss operates on the predicted absolute coordinates and combines four components. The first is the root mean squared distance computed in the original receptor coordinate frame, without any alignment, which directly supervises where the ligand sits in the binding site. The second is a mean squared error computed after optimally aligning the prediction to the reference using the classical Kabsch algorithm, implemented differentiably so that gradients flow through the singular value decomposition; this auxiliary term evaluates structural agreement after removing the best rigid transformation. The third term regularizes the complete pairwise distance matrices of the predicted and reference ligand structures, preserving internal ligand geometry independently of absolute position. The fourth is a clash penalty that quadratically punishes protein-ligand atom pairs closer than a cutoff of 2.0 in the normalized coordinate representation. A sensitivity analysis settled on weights of 1.0, 0.10, 0.05, and 0.02 for these four terms, respectively.

That sensitivity analysis produced one of the study’s most instructive findings. With only the receptor-frame RMSD term active, the dual-branch model achieved a mean RMSD of 5.08 angstroms on the PDBbind-2020 benchmark. Adding the auxiliary terms lowered this to 4.67 angstroms, a clear demonstration that alignment-aware, intraligand, and clash supervision genuinely help. But the reverse experiment was even more striking: when the weight of the raw RMSD term was reduced while the auxiliary terms stayed fixed, performance collapsed catastrophically, from 4.67 angstroms at full weight to 27.96 angstroms when the term was removed entirely. The auxiliary geometric terms complement rather than replace the supervision needed for correct absolute placement, a nuance the authors argue is essential for anyone designing docking losses.

The team evaluated seven graph neural network configurations within the framework, including TransformerConv, GraphSAGE, Graph Attention Networks, ARMAConv, and three hybrid architectures combining GIN, GAT, and TransformerConv layers. TransformerConv, an attention-based message-passing operator that can incorporate edge attributes, delivered the strongest overall performance, followed by GraphSAGE and the G3 hybrid, though it also required the most training time per epoch. All experiments were run on modest hardware, an Intel Core i7-7700 desktop with a GeForce GTX 1080 GPU, and five training runs with different random seeds showed a standard deviation of only about 0.01 angstroms, indicating that the results are highly reproducible.

The headline comparison, against MedusaGraph under strictly identical conditions, same initial MedusaDock poses, same preprocessing, same test complexes, atom ordering, graph construction, and RMSD calculation, showed consistent gains. On PDBbind-2020, a dataset of 3,956 preprocessed protein-ligand complexes split by CD-HIT sequence clustering at 90 percent identity, MedusaGraph achieved a mean RMSD of 5.08 angstroms, while KTransPose reached 4.67, 4.51, and 4.47 angstroms across three successive refinement iterations. On the independent CASF-2016 benchmark of 278 evaluated complexes, MedusaGraph’s 4.81 angstroms fell to 4.41, 4.27, and 4.24 angstroms for KTransPose, an improvement of roughly 12 percent on both datasets. Paired two-sided Wilcoxon signed-rank tests confirmed the differences were statistically significant at the matched iterations. Notably, iterative refinement for KTransPose reduced RMSD monotonically, whereas MedusaGraph actually degraded at its third iteration on both datasets, rising from 4.68 to 4.90 angstroms on PDBbind-2020.

The refinement dynamics themselves carry a practical lesson. Most of the total improvement over the initial MedusaDock pose is achieved in the very first iteration, with the second and third iterations contributing progressively smaller gains, just 0.16 and then 0.04 angstroms for the TransformerConv model on PDBbind-2020. Because each additional iteration requires applying the preceding frozen models, updating coordinates, and reconstructing the ligand-dependent graph edges within a 6 angstrom distance threshold, the authors stopped at three iterations, judging the diminishing returns not worth the extra computational cost. Success-rate analyses at fixed RMSD thresholds added further texture: KTransPose showed its clearest advantages at the 3 and 5 angstrom cutoffs, while the strict 2 angstrom success rate remained low for both methods and was not uniformly improved. A supplementary analysis using the DAAP affinity predictor also found that KTransPose-refined complexes yielded better affinity-prediction metrics across Pearson correlation, error, and concordance measures than MedusaGraph outputs.

The authors are candid about the boundaries of their results. KTransPose refines an existing docked pose rather than performing blind or de novo docking, so its performance depends on the quality of the supplied MedusaDock starting point, and protein coordinates are held fixed throughout, meaning receptor flexibility and conformational change are not modeled. The comparison with MedusaGraph is a method-level comparison, so the gains cannot be attributed to the global branch or the loss function alone, and the evaluation rests primarily on PDBbind-2020 with CASF-2016 as an additional test. Even so, the framework’s consistent improvements, its reproducibility, and its publicly released codebase mark it as a meaningful step for computational drug discovery, and the authors point toward equivariant architectures, flexible-receptor settings, and more diverse docking initializations as the next frontiers. For a field where fractions of an angstrom can separate a plausible drug candidate from a dead end, a reliable 12 percent reduction in pose error is no small matter.

Subject of Research: Deep learning-based refinement of protein-ligand binding pose prediction in structure-based drug discovery

Article Title: Enhancing protein–ligand pose prediction via dual-branch deep learning with docking-oriented multi-component loss and iterative refinement

Article References: Alam, M. K., Rahman, J., Newton, M. A. H., & Ali, M. E. (2026). Enhancing protein–ligand pose prediction via dual-branch deep learning with docking-oriented multi-component loss and iterative refinement. Applied Intelligence, 56(15), Article 458. https://doi.org/10.1007/s10489-026-07475-9

Image Credits: AI Generated

DOI: 10.1007/s10489-026-07475-9

Keywords: protein-ligand pose prediction, molecular docking, graph neural networks, drug discovery, KTransPose, MedusaGraph, PDBbind-2020, CASF-2016, Kabsch alignment, TransformerConv, iterative refinement, virtual screening

The biological target at the heart of the study is a pair of enzymes known as cyclin-dependent kinases 4 and 6, or CDK4 and CDK6. These kinases act as gatekeepers of the cell division cycle, driving the transition from the G1 phase, when cells grow and prepare, into S phase, when DNA is replicated. When the CDK4/6 machinery is dysregulated, cells can escape normal growth controls and proliferate uncontrollably, a hallmark of many malignancies. The importance of these enzymes is underscored by the clinical success of approved CDK4/6 inhibitors such as palbociclib, ribociclib and abemaciclib, which have transformed the treatment landscape for hormone receptor-positive breast cancer and are being explored in other tumor types, including lung cancer.

Yet the existing drugs are not the end of the story. Resistance mechanisms emerge, side effects limit some patients, and clinicians and researchers continue to call for chemically diverse scaffolds, meaning molecules with fundamentally different structural backbones, that can potently inhibit both CDK4 and CDK6 at the same time. Finding such dual inhibitors by traditional high-throughput screening is expensive and slow, which is precisely why the research team turned to computer-aided drug discovery. Their strategy was to build a computational funnel: a wide opening that could swallow an entire chemical library and progressively narrower stages that would discard weak candidates until only the most promising molecules remained.

The first stage of the funnel was ligand-based machine learning. Rather than requiring knowledge of the target protein’s three-dimensional structure, ligand-based models learn from the known activities of compounds that have already been tested. The researchers assembled curated datasets from the ChEMBL database, a public repository of bioactivity data, gathering 265 compounds with measured activity against CDK4 and 402 against CDK6. Each molecule was converted into a numerical representation called an extended-connectivity fingerprint, specifically the ECFP4 format, which encodes the local atomic environments of a molecule in a way that machine learning algorithms can digest. Fingerprint-based descriptors of this kind have become a workhorse of computational chemistry because they capture subtle structural features that often correlate with biological activity.

With the data prepared, the team benchmarked multiple machine learning algorithms to see which could best predict inhibitory potency. The winner was a Bayesian Ridge regressor, a statistical learning method that combines ridge regression with Bayesian regularization, making it robust against overfitting on modestly sized datasets. Trained on the ECFP4 fingerprints, the model achieved cross-validated coefficients of determination, R-squared values, of 0.731 plus or minus 0.022 for CDK4 and 0.721 plus or minus 0.070 for CDK6. In practical terms, this means the model could explain roughly seventy percent of the variance in compound potency from molecular structure alone, a level of predictive accuracy strong enough to make meaningful prioritization decisions. The consistency of performance across both targets was crucial, because the goal was a dual inhibitor, and a model that excelled at one kinase but stumbled on the other would have undermined the entire approach.

The trained machine learning filter was then deployed against a library of 22,823 compounds. The Bayesian Ridge model scored every molecule for its likely activity against CDK4 and CDK6, allowing the researchers to discard the vast majority of the library computationally before any expensive calculation was run. This is where the economics of the approach become compelling: instead of assaying or docking tens of thousands of compounds, the team could concentrate structural modeling resources on the small fraction that the statistical model deemed most promising. The survivors of the machine learning filter then entered the second stage of the funnel, molecular docking against both CDK4 and CDK6.

Docking is a structure-based technique that uses three-dimensional models of the target proteins, typically derived from experimentally determined structures in the Protein Data Bank, to predict how a small molecule physically fits into the binding pocket of the enzyme. The docking calculations, performed at standard precision, evaluated the geometric complementarity and predicted binding energy of each candidate within the ATP-binding sites of both kinases. By requiring candidates to dock well against both CDK4 and CDK6, the researchers enforced dual-target affinity at the structural level, complementing the statistical predictions of the machine learning stage. This dual-target docking refinement whittled the prioritized list down to three candidate hits selected for biochemical evaluation in the laboratory.

The experimental results validated the computational strategy emphatically. Of the three candidates tested, HY-18,623 emerged as a potent dual inhibitor, with a half-maximal inhibitory concentration, IC50, of 3.5 nanomolar against CDK4 and 17.4 nanomolar against CDK6. Values in the low nanomolar range represent the kind of potency typically associated with advanced lead compounds, and achieving it against both kinases simultaneously with a single molecule is exactly what the field has been seeking. The result suggests that the computational funnel did not merely enrich for plausible binders but genuinely identified a molecule with drug-development potential.

To understand why HY-18,623 binds so tightly, the researchers turned to the third pillar of their workflow: molecular dynamics simulations. Docking provides a static snapshot of a predicted binding pose, but proteins and ligands in solution are constantly in motion, and a pose that looks perfect in a single frame may fall apart within nanoseconds. The team ran 200-nanosecond molecular dynamics simulations of the CDK4 and CDK6 complexes with HY-18,623, tracking the stability of the bound state over time using metrics such as root mean square deviation of the atomic positions. The simulations revealed that the compound maintains persistent hydrogen bonds with the so-called hinge residues of both kinases, Val96 in CDK4 and Val101 in CDK6. Hinge-region hydrogen bonding is a canonical anchor of kinase inhibitor binding, and its persistence throughout the simulations indicates a stable, high-affinity binding mode. Binding free energy analyses further quantified the favorable thermodynamics of the interaction, corroborating the experimental potency measurements.

Beyond the specific discovery of HY-18,623, the study offers a template for how modern drug discovery campaigns can be organized. The sequential integration of ligand-based machine learning, structure-based docking and dynamics-based validation creates a hierarchy of evidence, in which each stage interrogates the candidates from a different angle: statistical structure-activity relationships, physical fit within the binding site, and dynamic stability of the complex. The researchers have also made their data publicly available through a GitHub repository, supporting the open-science ethos that is increasingly shaping kinase drug discovery. For a field in which approved CDK4/6 inhibitors have already changed the course of breast cancer treatment, yet resistance and toxicity continue to motivate the search for alternatives, the arrival of a computationally discovered, nanomolar dual inhibitor with a structurally characterized binding mode is a development worth watching. HY-18,623 now stands as a promising lead candidate for further therapeutic development in oncology, and the workflow that produced it as a potentially reusable engine for discovering the next generation of kinase inhibitors.

Subject of Research: Computational virtual screening combining machine learning, docking and molecular dynamics to discover dual CDK4/6 inhibitors for cancer therapy

Article Title: Harnessing machine learning, docking and molecular dynamics for the virtual screening of compounds as CDK4/6 dual inhibitors

Article References: Wang, Y., Fang, L., Liu, Q., Zhang, Z., Xu, C., & Jiang, Y. (2026). Harnessing machine learning, docking and molecular dynamics for the virtual screening of compounds as CDK4/6 dual inhibitors. Molecular Diversity. https://doi.org/10.1007/s11030-026-11716-x

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11716-x

Keywords: CDK4, CDK6, virtual screening, machine learning, molecular docking, molecular dynamics, drug discovery, kinase inhibitors, cancer, ChEMBL, ECFP4 fingerprints, HY-18,623

The biological target at the heart of the study is one of the most heavily exploited in modern oncology. VEGF signalling is the master switch of tumour angiogenesis, the process by which cancers recruit their own blood supply, and blocking it has produced some of the first triumphs of targeted therapy. The monoclonal antibody bevacizumab, approved by the FDA in 2004 for metastatic colorectal cancer, starves tumours by intercepting VEGF itself, while a growing family of small-molecule tyrosine kinase inhibitors including sorafenib, sunitinib, pazopanib, regorafenib, cabozantinib, axitinib, tivozanib, vandetanib, lenvatinib and nintedanib jam the intracellular signalling machinery of VEGFR-2. Yet these drugs have been most successful in liver, gastrointestinal, renal and thyroid malignancies, and their efficacy in brain tumours has remained stubbornly disappointing, leaving a clear opening for better molecules.

Glioblastoma multiforme is the context in which that unmet need is starkest. It is the most common malignant brain tumour, accounting for nearly half of all such cases, and it carries one of the grimmest prognoses in medicine. The current standard of care combines surgery, radiotherapy and the DNA-methylating agent temozolomide, supplemented in some settings by nitrosoureas such as lomustine and carmustine that cross the blood-brain barrier. All of these carry serious toxicities, from haematological suppression to pulmonary and ocular damage. Because glioblastoma neovascularisation is heavily regulated through VEGF and VEGFR-2, the receptor remains an attractive target, and one published patent has already proposed VEGFR-2 blockade as a treatment strategy for sensitive cancers. What has been missing is a reliable way to find new chemical matter for the purpose.

The Tartu researchers began with the VEGFR-2 binding site itself, which presents four pharmacophoric regions that successful inhibitors must engage. A heteroaromatic ring occupies the hinge region of the catalytic ATP binding pocket; a hydrophobic aryl or heteroaryl group fills the space leading toward the Asp-Phe-Gly motif; a hydrogen-bond donor and acceptor linker such as a urea or amine addresses the DFG domain; and, when the receptor adopts its inactive conformation, a deep allosteric hydrophobic pocket opens that confers selectivity. Rather than modelling a single congeneric series, as most prior QSAR efforts have done, the team assembled a dataset of 216 VEGFR-2 inhibitors spanning twelve distinct parent scaffolds, including phthalazines, pyridines, quinazolines, quinoxalines, benzoxazoles, benzothiazoles, coumarins, thiazolidine-2,4-diones, thioureas, nicotinamides, quinolines and isatins, all manually curated from a decade of literature with careful attention to assay consistency, duplicates, salts and isomers.

From this dataset, divided into 173 training and 43 test compounds, the researchers built a five-parameter quantitative structure-activity relationship model using best multiple linear regression with step-forward descriptor selection. The statistical performance was impressive for such a chemically heterogeneous training pool: a coefficient of determination of 0.824 on the training set, a cross-validated value of 0.807, and an external test set correlation of 0.74, with concordance correlation coefficients above 0.85 in both settings. A thousand-step Y-randomisation returned an essentially zero coefficient of determination, confirming that the model captured genuine structure-activity signal rather than chance correlation. Critically, all five descriptors are interpretable. The maximum carbon electrotopological state and the polarizability-weighted Geary autocorrelation contribute positively to activity, the mass-weighted Broto-Moreau autocorrelation and ionisation-potential-weighted Geary term contribute negatively, and a summed nitrogen-hydrogen E-state index rewards the presence of amino linker motifs, mirroring the pharmacophore logic of the binding site.

The model was then turned loose on chemical databases. Drawing on the structure-activity analysis of the most active phthalazine, pyridine and quinazoline derivatives, the team designed five extended scaffolds, or chemotypes, including pyridine variants bearing phenyl substituents and a phthalazine in which a naphthalene moiety, known to confer potent sub-nanomolar VEGFR-2 inhibition, was attached through an amino linker. Screening the MolPort and ZINC15 databases with these chemotypes and ranking candidates by predicted pIC50 yielded eight purchasable compounds with predicted activities between 8.4 and 9.5, all within the model’s applicability domain. Each was then purchased and put to the test in a cell-free VEGFR-2 kinase assay.

One molecule stood out. Compound G, formally N-(naphthalen-2-yl)-4-[4-(prop-2-yn-1-yloxy)phenyl]phthalazin-1-amine, inhibited VEGFR-2 with an IC50 of 0.497 micromolar, verified by high-performance liquid chromatography purification to better than 98 percent purity and confirmed across three independent experiments. Its naphthalene-bearing phthalazine architecture, the researchers note, has no close counterpart among reported VEGFR-2 inhibitors in PubChem, ZINC15 or ChEMBL, making it a genuinely novel chemotype and a credible analogue of the clinical candidate vatalanib. The other seven compounds, despite structural similarity to known actives and confident predictions, fell short, including a methoxy-substituted pyridine that differed from an active literature molecule by a single methyl ether group. That mixed outcome is itself informative, a candid illustration of the gap between prediction and biology.

The cellular results added nuance. In MTT viability assays after 72 hours of treatment, compound G reduced viability of H4 neuroglioma cells with an IC50 of 14.4 micromolar and A172 glioblastoma cells with an IC50 of 27.3 micromolar, both in a concentration-dependent manner. The non-malignant HEK293 reference model showed no significant loss of viability, and the breast cancer line MCF-7 was only marginally affected, hinting at selectivity for glial tumour cells. However, the high-grade U118MG glioblastoma line did not respond significantly, and the authors are careful to explain why: enzyme inhibition in a test tube does not guarantee cellular efficacy, which depends on compound permeability, intracellular target access and compensatory survival pathways. These findings echo earlier work showing that selective VEGFR-2 blockade suppresses VEGF-driven growth in astrocytoma models, while also reflecting the known limitations of anti-angiogenic drugs, which rarely kill tumour cells directly and often deliver modest survival benefits.

The authors position the new phthalazine hit as a starting point rather than a drug. It now requires optimisation, confirmation of target engagement inside cells, ADMET profiling and broader biological validation, and the team argues that pairing VEGFR-2 blockade with inhibition of EGFR, c-Met, BRAF or epigenetic targets such as HDAC could overcome the resistance and vessel co-option that blunt single-target anti-angiogenic therapy. What makes the study a compelling template is its transparency: the QSAR models and curated data are publicly deposited in the QsarDB repository with a digital object identifier, allowing other groups to interrogate, reuse and extend the work. At a moment when artificial intelligence in drug discovery is often oversold, this pipeline, modest dataset, explainable model, honest experimental accounting and a verified novel hit against one of medicine’s most feared cancers, shows what careful, verifiable machine learning can actually deliver.

Subject of Research: Machine learning-guided discovery of a novel phthalazine-based VEGFR-2 inhibitor evaluated in glioblastoma cell lines

Article Title: Discovery of a novel phthalazine-based VEGFR-2 inhibitor by machine learning analysis of diverse chemical space with favourable pharmacophores and its evaluation in glioblastoma cell lines

Article References: Zukic, S., Ivanova, L., Zusinaite, E., Merits, A., & Maran, U. (2026). Discovery of a novel phthalazine-based VEGFR-2 inhibitor by machine learning analysis of diverse chemical space with favourable pharmacophores and its evaluation in glioblastoma cell lines. Results in Chemistry, 30, Article 103884. https://doi.org/10.1016/j.rechem.2026.103884

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.103884

Keywords: VEGFR-2, phthalazine, machine learning, QSAR, glioblastoma, angiogenesis, virtual screening, drug discovery, tyrosine kinase inhibitor, naphthalene, cancer therapy, computational chemistry

The team began with the InterBioscreen (IBS) database, a curated collection of 69,075 natural products and natural product-like compounds. Such libraries are attractive starting points for antiviral drug discovery because natural scaffolds have been shaped by evolution to interact with proteins, often displaying chemical complexity and stereochemistry that synthetic combinatorial libraries struggle to match. Rather than docking every molecule against an arbitrary structure, the researchers employed a structure-guided strategy centered on the SARS-CoV-2 main protease, also known as Mpro or 3CL protease, a cysteine protease that cleaves the viral polyproteins at eleven conserved sites and is essential for viral replication. Because the enzyme has no close human homolog, inhibitors are less likely to interfere with host physiology, a property that has made Mpro one of the most thoroughly validated antiviral targets since the pandemic began.

High-throughput virtual screening of the full library produced a ranked list of candidate binders, which the team then subjected to more stringent re-docking to confirm that the apparent binding poses were reproducible and geometrically sensible. From this refined analysis, seven compounds emerged as hits. Each paired a high predicted binding affinity with specific hydrogen-bond interactions to residues lining the enzyme’s active site, the catalytic pocket where the virus cuts its own polyproteins. Notably, all seven compounds were new to this target and to the virus itself, meaning the screen had not simply rediscovered previously known inhibitors but surfaced an entirely fresh set of chemical scaffolds for evaluation.

Docking scores alone are a notoriously unreliable predictor of true binding, so the researchers advanced all seven hits to molecular dynamics (MD) simulations. These simulations solvate the protein-ligand complexes in explicit water and let the system evolve under physical force fields, revealing whether a docked pose survives the constant thermal jostling of a realistic molecular environment or falls apart within nanoseconds. The stability of each complex was further quantified using the MM-GBSA method, which estimates binding free energy by combining molecular mechanics interaction energies with continuum solvent models and can be averaged across many simulation snapshots to rank ligands more meaningfully than a single docking score. In parallel, the team predicted the ADME/Tox profiles, the absorption, distribution, metabolism, excretion, and toxicity characteristics, of each compound computationally, filtering out molecules likely to fail on pharmacokinetic grounds before any wet-lab work began.

After the dynamics-based triage, only three of the original seven compounds, STOCK1N-86169, STOCK1N-88750, and STOCK1N-65657, remained as high-potential candidates for Mpro inhibition. These three then entered the laboratory. The team ran a FRET-based Mpro inhibition assay, a standard enzymatic test in which a fluorescent peptide substrate is cleaved by the protease and inhibition is read as a preserved fluorescence signal, and an ELISA-based SARS-CoV-2 inhibition assay to probe antiviral activity more directly. The results were unambiguous: STOCK1N-86169 was the only potent inhibitor of the trio, inhibiting the enzyme with an IC50 of 3.8 micromolar, a concentration comfortably within the range expected of a chemical starting point for medicinal chemistry optimization.

Binding was independently confirmed by differential scanning calorimetry (DSC), a thermodynamic technique that measures the temperature at which a protein unfolds. Ligand binding typically stabilizes a protein against thermal denaturation, and the magnitude of the shift in melting temperature serves as a label-free readout of binding strength. When STOCK1N-86169 was bound to Mpro, the melting temperature of the complex rose by approximately 15 degrees Celsius relative to the unliganded protein, a substantial increase that corroborated the enzymatic assay and confirmed that the compound genuinely engages the protease rather than interfering artifactually with the assay chemistry.

The antiviral picture in cells was more nuanced. In cytotoxicity assays, STOCK1N-86169 showed a CC50 greater than 100 micromolar in HeLa cells, roughly 100 micromolar in Huh-7 human liver cells, and about 33.5 micromolar in Vero E6 kidney cells, indicating a moderate degree of cell-type variation in tolerance. In a Vero cell-based SARS-CoV-2 infection assay, the compound suppressed viral replication with an EC50 of approximately 7 micromolar, yielding a selective index, the ratio of cytotoxic to antiviral concentration, of around fivefold. A selective index of that size is modest by the standards of a clinical candidate, but for a first hit from an unoptimized natural product-like scaffold it represents a credible foundation, and the authors suggest the compound may be a promising candidate for further development as an effective inhibitor of SARS-CoV-2 Mpro.

The study sits within a broader and intensely active effort to expand the chemical arsenal against the main protease. Early in the pandemic, structures of Mpro published in 2020 enabled a global wave of inhibitor discovery, and since then the field has explored covalent peptidomimetics modeled on the enzyme’s substrate, non-covalent small molecules, and repurposed drugs. Plant-derived polyphenols such as baicalein, quercetin, and rutin have been reported as low-micromolar Mpro inhibitors, and open-science initiatives have iteratively refined screening hits into leads with improved drug-like properties. Against this backdrop, the new work is distinguished by its systematic use of a natural product-focused commercial library and by the completeness of its validation chain, which runs from virtual screening through dynamics, enzymology, thermodynamics, and live-virus assays, a chain that many purely computational studies never complete.

The methodological discipline on display is itself instructive. The pipeline incorporated several widely recommended safeguards: re-docking to filter initial screening artifacts, MD simulation to eliminate unstable poses, MM-GBSA energetics to rank complexes quantitatively, and ADME/Tox prediction to remove molecules with obvious pharmacokinetic liabilities. The design philosophy echoes lessons learned from decades of virtual screening, including the need to beware of pan-assay interference compounds, molecules that masquerade as inhibitors by aggregating or fluorescing nonspecifically. By demanding convergence between computational predictions and two independent experimental readouts, enzymatic inhibition and thermal stabilization, before declaring a winner, the Kolkata and Chandigarh team avoided many of the false positives that plague database-driven antiviral screens.

Whether STOCK1N-86169 ultimately advances toward a drug will depend on the usual gauntlet of lead optimization: improving potency, clarifying its mechanism of binding at atomic resolution, tuning its selectivity index upward, and establishing pharmacokinetic behavior in animals. Its novelty as a scaffold, unprecedented for this target and for the virus according to the authors, is an asset, since new chemotypes give medicinal chemists fresh territory for structural modification when resistance or tolerability problems arise with existing inhibitor classes. For now, the study stands as a template for rigorous antiviral hit discovery, showing that a carefully filtered virtual screen of a natural product database, when anchored to experimental verification at every stage, can still deliver genuinely new chemical matter against one of the most heavily mined targets in modern virology.

Subject of Research: Discovery of a natural product-like inhibitor of the SARS-CoV-2 main protease through virtual screening and in vitro validation

Article Title: Discovery of a new SARS-CoV-2 Mpro inhibitor from a natural product and natural product-like compound database using in silico and in vitro studies

Article References: Das, S., Pal, U., Begum, Y., Joshi, A., Singh, N., Thakur, K. G., & Singh, U. P. (2026). Discovery of a new SARS-CoV-2 Mpro inhibitor from a natural product and natural product-like compound database using in silico and in vitro studies. Molecular Diversity. https://doi.org/10.1007/s11030-026-11724-x

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11724-x

Keywords: SARS-CoV-2, Mpro, main protease inhibitor, virtual screening, natural products, molecular dynamics, MM-GBSA, FRET assay, differential scanning calorimetry, antiviral drug discovery, InterBioscreen, ADME prediction

The logic of the target is compelling. GSK-3β is a constitutively active serine/threonine kinase that phosphorylates more than 100 substrates, governing glucose disposal, lipid metabolism, inflammation and fibrogenesis. Under normal conditions, insulin signalling activates the PI3K/Akt cascade, which phosphorylates GSK-3β at Ser9 and switches it off, freeing glycogen synthase to build glycogen stores in liver and muscle. In insulin resistance—the shared root of both type 2 diabetes and MASH—that brake fails. Hyperactive GSK-3β keeps glycogen synthase locked in its inactive, phosphorylated state, degrades insulin receptor substrate-1 through inhibitory serine phosphorylation, stabilizes the gluconeogenic factors PGC-1α and FOXO1, and drives hepatic glucose output. In the liver, it fuels SREBP-1c-mediated lipogenesis, undermines AMPK-driven fatty acid oxidation, sustains NF-κB inflammatory signalling and promotes TGF-β/Smad-driven collagen deposition in hepatic stellate cells. Preclinical rodent studies have repeatedly shown that inhibiting or deleting GSK-3β improves insulin sensitivity and blunts steatosis, inflammation and fibrosis.

Existing GSK-3 inhibitors have not translated into metabolic medicine. Lithium is non-selective with a narrow therapeutic window; AR-A014418 remains a biochemical tool compound; tideglusib, a clinical candidate for neurodegenerative disease, crosses the blood–brain barrier—an undesirable trait for chronic peripheral therapy—and LY2090314 and elraglusib have been pursued in oncology. Shared obstacles include poor selectivity between the GSK-3α and GSK-3β isoforms, whose ATP sites are nearly identical, and the theoretical oncogenic risk of sustained β-catenin stabilization. These gaps motivated the research team, led by Lateef Bello and colleagues at Adekunle Ajasin University in Nigeria, to search for peripherally restricted, structurally novel scaffolds in microbial chemistry.

The screening cascade began with 36,545 microbial natural products from the Natural Products Atlas. Applying Lipinski’s Rule of Five in DataWarrior retained 21,390 drug-like compounds—58.5 percent of the library—suggesting that microbial metabolites occupy an unusually favourable region of chemical space. A four-feature energy-optimized pharmacophore, derived from the crystal structure of human GSK-3β bound to the ATP-competitive inhibitor 7YG (PDB ID 4ACC), then filtered the library down to just 218 compounds matching three aromatic ring features and one hydrogen-bond donor. This ligand-based step eliminated 99 percent of candidates before any computationally expensive docking was attempted.

Critical to the study’s credibility was rigorous protocol validation. Re-docking the co-crystallized ligand reproduced the crystallographic pose with a root mean square deviation of 0.8357 angstroms, well within the conventional 2.0 angstrom cutoff. Enrichment analysis using 25 known GSK-3β inhibitors from ChEMBL and 1,255 generated decoys demonstrated exceptional discriminatory power: a ROC value of 0.98, an area under the curve of 0.97, and full recovery of all active compounds within the top 5 percent of the ranked list—a twenty-fold enrichment over random selection.

Structure-based docking then proceeded through three escalating precision tiers. High-throughput virtual screening cut the 218 pharmacophore hits to 178; standard precision docking at a −6.0 kcal/mol threshold retained 88 compounds, and a −6.5 kcal/mol cut kept 69. Extra-precision (XP) docking of the 50 best-ranked compounds ultimately yielded six hits—NPA011425, NPA014875, NPA020257, NPA004276, NPA034843 and NPA036064—with XP scores ranging from −6.858 to −11.059 kcal/mol. Every one of them outperformed the co-crystallized reference ligand (−6.455 kcal/mol), AR-A014418 (−5.929 kcal/mol) and tideglusib (−5.895 kcal/mol). All six formed hydrogen bonds with the conserved hinge-region residues Asp133, Val135 and Pro136 while packing against the hydrophobic ATP-binding pocket lined by Ile62, Val70, Ala83, Leu132 and Leu188, a signature consistent with ATP-competitive inhibition.

Independent scoring methods corroborated the docking results. Molecular Mechanics Generalized Born Surface Area (MMGBSA) calculations gave NPA014875 a binding free energy of −62.67 kcal/mol, essentially matching the native ligand’s −63.15 kcal/mol and far exceeding the reference inhibitors. Induced Fit Docking, which permits the binding site to relax around each ligand, confirmed receptor-adaptive binding: NPA014875 scored −730.74 kcal/mol, comparable to AR-A014418 and superior to both the reference ligand and tideglusib, while several compounds gained new interactions—such as π–π stacking with Phe67 or cationic contacts with Arg141—once side-chain flexibility was introduced. A machine-learning AutoQSAR model, trained on 1,500 curated ChEMBL inhibitors with an R-squared of 0.7123, predicted the highest potency for NPA020257, with a pIC50 of 6.82.

The pharmacokinetic and safety picture was equally encouraging. All six hits showed zero Lipinski violations, molecular weights between 308 and 430 g/mol, and favourable oral bioavailability scores. Crucially, none was predicted to cross the blood–brain barrier, whereas tideglusib was—a property the authors highlight as a therapeutic advantage, since central GSK-3β inhibition has been linked to neuropsychiatric effects and circadian disruption. Four of the six compounds showed no predicted inhibition across the five major cytochrome P450 isoforms, an important consideration for diabetic patients on polypharmacy. Toxicity predictions with ProTox-3.0 assigned all hits to class IV, with none predicted cytotoxic or cardiotoxic. The only flags were specific: NPA011425 alone was predicted hepatotoxic, NPA004276 showed oestrogen-receptor activity, and NPA014875 and NPA004276 were predicted active in mitochondrial membrane potential assays—a liability that matters in MASH, where hepatocyte mitochondrial reserve is already compromised.

Reading all five evaluation axes together—docking score, binding free energy, induced-fit performance, predicted potency and toxicity liabilities—the study names NPA020257 and NPA034843 as the most balanced candidates, both carrying zero major predicted liabilities. NPA014875 stands out as the strongest calculated binder but requires lead optimization against its mitochondrial flag; NPA011425 and NPA004276 remain affinity-rich chemotypes with clearly defined medicinal-chemistry objectives; and NPA036064 offers the cleanest safety profile as a backup scaffold. The authors are careful to frame the composite ranking as a triage map rather than a verdict, keeping all six compounds in play.

The limitations are candidly acknowledged. Every finding is computational and awaits experimental confirmation: enzymatic GSK-3β assays, kinase selectivity profiling against GSK-3α and broader panels, molecular dynamics simulations of complex stability, and validation of the predicted ADMET profiles in cellular and animal models of diabetes and MASH. The AutoQSAR model was trained largely on synthetic inhibitors, and no formal applicability-domain check was performed for the natural product hits. Still, the work demonstrates the power of a disciplined virtual screening cascade to convert a massive natural product database into a short, chemically diverse list of prioritized hits—hits that, if validated, could one day deliver what current medicine cannot: a single molecule that simultaneously restores insulin sensitivity and halts the inflammatory, fibrotic progression of fatty liver disease.

Subject of Research: Computational identification of microbial natural product inhibitors of glycogen synthase kinase 3 beta as candidate dual therapeutics for type 2 diabetes mellitus and MASH.

Article Title: Computational screening of the natural products atlas identifies potential glycogen synthase kinase 3 beta inhibitors as candidate therapeutics for type 2 diabetes mellitus and MASH

Article References: Bello, L., Nwankwo, D. O., Shodehinde, S. A., Akerele, G. P., Okuntimehin, B., Ogunjobi, E. G., Awelewa, O. V., Abass, O. A., Oginni, S. A., & Olubode, S. O. (2026). Computational screening of the natural products atlas identifies potential glycogen synthase kinase 3 beta inhibitors as candidate therapeutics for type 2 diabetes mellitus and MASH. Discover Chemistry, 3(1), Article 536. https://doi.org/10.1007/s44371-026-00989-8

Image Credits: AI Generated

DOI: 10.1007/s44371-026-00989-8

Keywords: GSK-3β, type 2 diabetes, MASH, Natural Products Atlas, virtual screening, molecular docking, MMGBSA, induced fit docking, AutoQSAR, ADMET, drug discovery, microbial natural products

The research team, led by The-Huan Tran, Thai-Son Tran and Thanh-Dao Tran of the University of Medicine and Pharmacy at Ho Chi Minh City and Hue University in Vietnam, constructed an integrative in silico workflow to mine a molecular library derived from known inhibitors of both enzymes, including the cholinesterase drug rivastigmine and the monoacylglycerol lipase inhibitors JZL-184 and ABX-1431. Acetylcholinesterase breaks down the neurotransmitter acetylcholine, and its inhibition has long underpinned symptomatic Alzheimer’s therapy. Monoacylglycerol lipase, by contrast, degrades 2-arachidonoylglycerol, an endocannabinoid signaling lipid whose preservation has been linked to dampened neuroinflammation and enhanced glial immunity. Hitting both enzymes with a single molecule could, in principle, address cognitive decline and inflammatory damage at the same time.

Starting from 365 candidate compounds, the investigators carried out molecular docking against crystal structures of human acetylcholinesterase and human monoglyceride lipase, using AutoDock Vina to predict binding poses and scores. They then applied interaction-based filtering to retain only those ligands that reproduced the key contacts characteristic of known inhibitors, followed by in silico prediction of absorption, distribution, metabolism and excretion properties and toxicity. This successive narrowing of the chemical space reflects a widely adopted paradigm in early-stage drug discovery, where computational filters are stacked so that only chemically sensible, pharmacokinetically plausible candidates advance to the most expensive and time-consuming analyses.

One compound, designated H34, rose to the top of the ranking. Molecular dynamics simulations run with GROMACS under the CHARMM36 force field examined the structural stability of the H34-enzyme complexes over time, tracking the root-mean-square deviation of the protein backbone, the root-mean-square fluctuations of individual residues, the radius of gyration and the solvent-accessible surface area. Across these metrics, the H34 complexes displayed comparatively favorable stability, suggesting that the ligand does not disrupt the overall fold of either enzyme and remains seated in the binding pocket under physiologically realistic conditions.

To move beyond qualitative stability assessments, the team estimated binding free energies using the MM/GBSA end-state method implemented in the gmx_MMPBSA tool. H34 yielded estimated binding free energies of −30.96 and −37.34 kcal/mol for the two enzyme targets, values that compare favorably within the context of the screened series and support the compound’s predicted affinity for both proteins. These calculations decompose the interaction into electrostatic, van der Waals and solvation contributions, offering a thermodynamic rationale for why the molecule holds on to each active site.

Further dynamical profiling added depth to the picture. ProLIF interaction mapping, which encodes protein-ligand contacts as fingerprints across the simulation trajectory, showed that the interactions anchoring H34 in each active site persisted over time rather than flickering in and out. Free energy landscape analysis, built on dihedral angle principal component analysis, mapped the conformational behavior of the complexes and revealed energetically stable basins, indicating that the ligand-bound enzymes do not wander between widely divergent conformations. Together, these analyses portray a compound whose binding mode is not merely a static docking artifact but a durable, low-energy arrangement.

Drug-likeness and safety filters reinforced the computational case for H34. SwissADME-based pharmacokinetic predictions indicated favorable properties relevant to oral absorption and brain exposure, while ProTox 3.0 predictions pointed to low acute toxicity. Because any candidate intended for Alzheimer’s disease must reach the central nervous system without accumulating liability elsewhere, these early pharmacokinetic and toxicological signals, though still predictive rather than experimental, are a meaningful part of the triage process that decides which molecules justify synthesis and laboratory testing.

The study’s authors are careful to frame H34 as a candidate rather than a drug. All findings derive from computational models, and docking scores, MM/GBSA estimates and simulations can only approximate the thermodynamics and kinetics of real enzyme inhibition. The necessary next steps include chemical synthesis and in vitro enzyme assays to measure actual inhibitory potency against both acetylcholinesterase and monoacylglycerol lipase, followed by cell-based and ultimately animal studies to test whether the dual-target hypothesis translates into cognitive and anti-inflammatory benefit in living systems.

Even so, the work illustrates why integrated computational strategies have become indispensable in multi-target drug discovery. By combining ligand-based library generation, structure-based docking, interaction fingerprinting, molecular dynamics, free energy estimation and ADME and toxicity prediction in a single pipeline, the researchers compressed an enormous search space into one well-characterized hit in silico. If H34 or its analogs survive experimental validation, the compound family could contribute to a new generation of multi-target-directed ligands designed for the complexity of Alzheimer’s disease, where modulating a single pathway has repeatedly fallen short.

Subject of Research: Computational discovery of dual acetylcholinesterase and monoacylglycerol lipase inhibitors for Alzheimer's disease

Article Title: Computational discovery and dynamic profiling of dual acetylcholinesterase and monoacylglycerol lipase inhibitors for Alzheimer’s disease

Article References: Tran, T.-H., Tran, T.-S., & Tran, T.-D. (2026). Computational discovery and dynamic profiling of dual acetylcholinesterase and monoacylglycerol lipase inhibitors for Alzheimer’s disease. Molecular Diversity. https://doi.org/10.1007/s11030-026-11725-w

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11725-w

Keywords: Alzheimer's disease, acetylcholinesterase, monoacylglycerol lipase, dual inhibition, molecular docking, molecular dynamics, MM/GBSA, virtual screening, drug discovery, neuroinflammation, multi-target ligands, ADME prediction

In a study published in Molecular Diversity, researchers led by Huang Zeng and Bo Qiu of Jiaying University describe a systematic computational framework that pairs five types of molecular fingerprints with both classical machine learning and deep learning algorithms. Their goal was twofold: to rigorously compare how different combinations of chemical representations and algorithms perform at predicting TOP1 inhibition, and to convert the best performers into practical tools that any medicinal chemist can use. The work culminated not just in models but in validated hit compounds, two of which showed meaningful inhibition of the enzyme in laboratory tests.

The foundation of the study is data. The team mined the ChEMBL database for compounds with experimentally measured TOP1 inhibitory activity, assembling a dataset suitable for training classification models. Each molecule was then encoded using five distinct fingerprint families: AtomPairs, which describe pairs of atoms and their topological separation; MACCS keys, a set of 166 structurally predefined substructure features; Morgan fingerprints, circular fingerprints that capture local atomic environments much like extended-connectivity fingerprints; PharmacoPFP, which encodes pharmacophore features rather than pure connectivity; and RDKit descriptors, a collection of calculated physicochemical and topological properties. Because each fingerprint represents a different view of what a molecule is, the choice of representation can matter as much as the choice of algorithm.

On top of these representations, the researchers built a total of forty models spanning classical methods such as support vector machines, random forests, and k-nearest neighbors, alongside deep neural networks. The head-to-head comparison produced a clear result: combinations built on Morgan fingerprints dominated the leaderboard. Under random data splitting, the four top performers were SVM paired with Morgan fingerprints, random forest with Morgan, a deep neural network using MACCS keys, and k-nearest neighbors with Morgan. All four achieved area under the receiver operating characteristic curve values, or ROC-AUC, between 0.93 and 0.94, a level of discrimination that places them firmly in the range useful for prioritizing compounds in virtual screening campaigns.

High scores alone can deceive, particularly in cheminformatics where models sometimes memorize dataset quirks rather than learning genuine structure-activity relationships. To guard against this, the team applied Y-scrambling, a validation technique in which the activity labels are randomly shuffled and models are retrained. If the original models had been learning noise, their scrambled counterparts would perform equally well. They did not, supporting the conclusion that the top models had captured real, non-random relationships between molecular structure and TOP1 inhibition. The authors are also candid about a second, more sobering test: when compounds were split by molecular scaffold rather than randomly, AUC values dropped to between 0.67 and 0.82. This scaffold-split performance reveals that the models are less reliable when asked to extrapolate to chemically novel frameworks, a well-known limitation of fingerprint-based approaches that the researchers explicitly quantify rather than hide.

Interpretability was another pillar of the work. Black-box predictions carry little weight in drug design unless chemists can see what drives them. Using SHAP analysis, a game-theoretic method that assigns each molecular feature a contribution to individual predictions, the team identified the key structural elements that the models associated with TOP1 inhibition. This step transforms the models from opaque ranking machines into sources of chemical hypotheses, allowing medicinal chemists to see which substructures and descriptor values push a compound toward or away from the predicted active class, and potentially to use those insights when designing next-generation analogs.

What distinguishes this study from many computational papers is its insistence on usability. The researchers deployed their best models as a freely accessible web application, reachable at drugpred.top:5000, where users can input a molecule as a SMILES string, the standard text notation for chemical structures, or simply draw the structure in a browser and receive a prediction of TOP1 inhibitory activity. For laboratories that cannot or prefer not to rely on an internet connection, standalone desktop executables are available for download on GitHub at github.com/zenghuang8006/TOP1-inhibitor-prediction. The authors note that earlier AI-driven studies of TOP1 inhibitors lacked both systematic fingerprint and algorithm comparisons and accessible predictive tools, and this pipeline is designed to fill both gaps at once.

The real proof, however, came from the laboratory. The team screened a commercial library of 189,554 compounds from SPECS using their computational framework, then took the highest-ranking candidates into in vitro testing. Two compounds, designated AG60 and AI61, emerged as potential TOP1 inhibitor hits. At a concentration of 400 micromolar, AG60 inhibited the enzyme by 64 percent, while AI61 achieved 90 percent inhibition. Converting a computational prediction into experimentally confirmed enzyme inhibition, even at moderate potency, is a meaningful step; the compounds now represent starting points, or hits, that can be chemically optimized into more potent and selective leads through medicinal chemistry.

The biological context underscores why this target continues to attract attention. TOP1 is essential in proliferating cells, and its inhibition selectively harms cancer cells that cannot tolerate the resulting DNA damage. Approved TOP1 poisons remain mainstays of treatment for colorectal, ovarian, and other cancers, and newer agents such as antibody-drug conjugates carrying TOP1 inhibitor payloads have expanded the target’s therapeutic reach. But resistance, toxicity, and the pharmacological limitations of existing scaffolds keep the demand for chemical novelty high. The authors’ own prior work, including the computational identification of FTY720 and COH29 as TOP1 catalytic inhibitors, illustrates a sustained research program aimed at expanding the toolbox of TOP1-directed chemistry.

For the broader field, the study offers a template as much as a set of hits. It demonstrates that fingerprint choice can be systematically benchmarked, that interpretability tools like SHAP can be integrated into the screening workflow, and that honest reporting of scaffold-split performance should accompany headline accuracy figures. It also shows that the distance between a machine learning model and a wet-lab result can be shortened when prediction tools are built for practicing chemists rather than for other data scientists. The forty models, the validated hits AG60 and AI61, and the freely available web and desktop predictors together form a practical framework that other target-focused screening campaigns can adapt. As artificial intelligence continues to compress the early stages of drug discovery, studies that pair rigorous benchmarking with experimental follow-through, and that are transparent about both strengths and limitations, offer the clearest picture of what these methods can actually deliver for anticancer drug development.

Subject of Research: AI-driven screening of topoisomerase I inhibitors using machine learning, deep learning and molecular fingerprints

Article Title: Integrating machine learning and deep learning with multiple molecular fingerprints for topoisomerase I inhibitor screening and lead identification

Article References: Integrating machine learning and deep learning with multiple molecular fingerprints for topoisomerase I inhibitor screening and lead identification. (n.d.). https://doi.org/10.1007/s11030-026-11709-w

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11709-w

Keywords: topoisomerase I, TOP1 inhibitors, machine learning, deep learning, molecular fingerprints, virtual screening, drug discovery, SHAP, ChEMBL, anticancer targets, SPECS compounds, QSAR