MetalProGNet: a structure-based deep graph model for metalloprotein-ligand interaction predictions

Interaction-Based Inductive Bias in Graph Neural Networks: Enhancing Protein-Ligand Binding Affinity Predictions From 3D Structures

Inductive bias in machine learning (ML) is the set of assumptions describing how a model makes predictions. Different ML-based methods for protein-ligand binding affinity (PLA) prediction have different inductive biases, leading to different levels of generalization capability and interpretability. Intuitively, the inductive bias of an ML-based model for PLA prediction should fit in with biological mechanisms relevant for binding to achieve good predictions with meaningful reasons. To this end, we propose an interaction-based inductive bias to restrict neural networks to functions relevant for binding with two assumptions: 1) A protein-ligand complex can be naturally expressed as a heterogeneous graph with covalent and non-covalent interactions; 2) The predicted PLA is the sum of pairwise atom-atom affinities determined by non-covalent interactions. The interaction-based inductive bias is embodied by an explainable heterogeneous interaction graph neural network (EHIGN) for explicitly modeling pairwise atom-atom interactions to predict PLA from 3D structures. Extensive experiments demonstrate that EHIGN achieves better generalization capability than other state-of-the-art ML-based baselines in PLA prediction and structure-based virtual screening. More importantly, comprehensive analyses of distance-affinity, pose-affinity, and substructure-affinity relations suggest that the interaction-based inductive bias can guide the model to learn atomic interactions that are consistent with physical reality. As a case study to demonstrate practical usefulness, our method is tested for predicting the efficacy of Nirmatrelvir against SARS-CoV-2 variants. EHIGN successfully recognizes the changes in the efficacy of Nirmatrelvir for different SARS-CoV-2 variants with meaningful reasons.

FragGen: towards 3D geometry reliable fragment-based molecular generation

3D structure-based molecular generation is a successful application of generative AI in drug discovery. Most earlier models follow an atom-wise paradigm, generating molecules with good docking scores but poor molecular properties (like synthesizability and drugability). In contrast, fragment-wise generation offers a promising alternative by assembling chemically viable fragments. However, the co-design of plausible chemical and geometrical structures is still challenging, as evidenced by existing models. To address this, we introduce the Deep Geometry Handling protocol, which decomposes the entire geometry into multiple sets of geometric variables, looking beyond model architecture design. Drawing from a newly defined six-category taxonomy, we propose FragGen, a novel hybrid strategy as the first geometry-reliable, fragment-wise molecular generation method. FragGen significantly enhances both the geometric quality and synthesizability of the generated molecules, overcoming major limitations of previous models. Moreover, FragGen has been successfully applied in real-world scenarios, notably in designing type II kinase inhibitors at the ∼nM level, establishing it as the first validated 3D fragment-based drug design algorithm. We believe that this concept-algorithm-application cycle will not only inspire researchers working on other geometry-centric tasks to move beyond architecture designs but also provide a solid example of how generative AI can be customized for drug design.

An Effective Plant Small Secretory Peptide Recognition Model Based on Feature Correction Strategy

Plant small secretory peptides (SSPs) play an important role in the regulation of biological processes in plants. Accurately predicting SSPs enables efficient exploration of their functions. Traditional experimental verification methods are very reliable and accurate, but they require expensive equipment and a lot of time. The method of machine learning speeds up the prediction process of SSPs, but the instability of feature extraction will also lead to further limitations of this type of method. Therefore, this paper proposes a new feature-correction-based model for SSP recognition in plants, abbreviated as SE-SSP. The model mainly includes the following three advantages: First, the use of transformer encoders can better reveal implicit features. Second, design a feature correction module suitable for sequences, named 2-D SENET, to adaptively adjust the features to obtain a more robust feature representation. Third, stack multiple linear modules to further dig out the deep information on the sample. At the same time, the training based on a contrastive learning strategy can alleviate the problem of sparse samples. We construct experiments on publicly available data sets, and the results verify that our model shows an excellent performance. The proposed model can be used as a convenient and effective SSP prediction tool in the future. Our data and code are publicly available at https://github.com/wrab12/SE-SSP/.

Assessing the performance of MM/PBSA and MM/GBSA methods. 10. Prediction reliability of binding affinities and binding poses for RNA-ligand complexes

Ribonucleic acid (RNA)-ligand interactions play a pivotal role in a wide spectrum of biological processes, ranging from protein biosynthesis to cellular reproduction. This recognition has prompted the broader acceptance of RNA as a viable candidate for drug targets. Delving into the atomic-scale understanding of RNA-ligand interactions holds paramount importance in unraveling intricate molecular mechanisms and further contributing to RNA-based drug discovery. Computational approaches, particularly molecular docking, offer an efficient way of predicting the interactions between RNA and small molecules. However, the accuracy and reliability of these predictions heavily depend on the performance of scoring functions (SFs). In contrast to the majority of SFs used in RNA-ligand docking, the end-point binding free energy calculation methods, such as molecular mechanics/generalized Born surface area (MM/GBSA) and molecular mechanics/Poisson Boltzmann surface area (MM/PBSA), stand as theoretically more rigorous approaches. Yet, the evaluation of their effectiveness in predicting both binding affinities and binding poses within RNA-ligand systems remains unexplored. This study first reported the performance of MM/PBSA and MM/GBSA with diverse solvation models, interior dielectric constants (εin) and force fields in the context of binding affinity prediction for 29 RNA-ligand complexes. MM/GBSA is based on short (5 ns) molecular dynamics (MD) simulations in an explicit solvent with the YIL force field; the GBGBn2 model with higher interior dielectric constant (εin = 12, 16 or 20) yields the best correlation (Rp = -0.513), which outperforms the best correlation (Rp = -0.317, rDock) offered by various docking programs. Then, the efficacy of MM/GBSA in identifying the near-native binding poses from the decoys was assessed based on 56 RNA-ligand complexes. However, it is evident that MM/GBSA has limitations in accurately predicting binding poses for RNA-ligand systems, particularly compared with notably proficient docking programs like rDock and PLANTS. The best top-1 success rate achieved by MM/GBSA rescoring is 39.3%, which falls below the best results given by docking programs (50%, PLNATS). This study represents the first evaluation of MM/PBSA and MM/GBSA for RNA-ligand systems and is expected to provide valuable insights into their successful application to RNA targets.

Unlocking Precision Docking for Metalloproteins

Metalloproteins play a fundamental role in molecular biology, contributing to various biological processes. However, the discovery of high-affinity ligands targeting metalloproteins has been delayed due, in part, to a lack of suitable tools and data. Molecular docking, a widely used technique for virtual screening of small-molecule ligand interactions with proteins, often faces challenges when applied to metalloproteins due to the particular nature of the ligand metal bond. To address these limitations associated with docking metalloproteins, we introduce a knowledge-driven docking approach known as "metalloprotein bias docking" (MBD), which extends the AutoDock Bias technique. We assembled a comprehensive data set of metalloprotein-ligand complexes from 15 different metalloprotein families, encompassing Ca, Co, Fe, Mg, Mn, and Zn metal ions. Subsequently, we conducted a performance analysis of our MBD method and compared it to the conventional docking (CD) program AutoDock4, applied to various metalloprotein targets within our data set. Our results demonstrate that MBD outperforms CD, significantly enhancing accuracy, selectivity, and precision in ligand pose prediction. Additionally, we observed a positive correlation between our predicted ligand free energies and the corresponding experimental values. These findings underscore the potential of MBD as a valuable tool for the effective exploration of metalloprotein-ligand interactions.

Screening of phytochemicals from Clerodendrum inerme (L.) Gaertn as potential anti-breast cancer compounds targeting EGFR: an in-silico approach

Breast cancer (BC) is the most prevalent malignancy among women around the world. The epidermal growth factor receptor (EGFR) is a tyrosine kinase receptor (RTK) of the ErbB/HER family. It is essential for triggering the cellular signaling cascades that control cell growth and survival. However, perturbations in EGFR signaling lead to cancer development and progression. Hence, EGFR is regarded as a prominent therapeutic target for breast cancer. Therefore, in the current investigation, EGFR was targeted with phytochemicals from Clerodendrum inerme (L.) Gaertn (C. inerme). A total of 121 phytochemicals identified by gas chromatography-mass spectrometry (GC-MS) analysis were screened against EGFR through molecular docking, ADMET analyses (Absorption, Distribution, Metabolism, Excretion, and Toxicity), PASS predictions, and molecular dynamics simulation, which revealed three potential hit compounds with CIDs 10586 [i.e. alpha-bisabolol (-6.4 kcal/mol)], 550281 [i.e. 2,(4,4-Trimethyl-3-hydroxymethyl-5a-(3-methyl-but-2-enyl)-cyclohexene) (-6.5 kcal/mol)], and 161271 [i.e. salvigenin (-7.4 kcal/mol)]. The FDA-approved drug gefitinib was used to compare the inhibitory effects of the phytochemicals. The top selected compounds exhibited good ADMET properties and obeyed Lipinski's rule of five (ROF). The molecular docking analysis showed that salvigenin was the best among the three compounds and formed bonds with the key residue Met 793. Furthermore, the molecular mechanics generalized born surface area (MMGBSA) calculations, molecular dynamics simulation, and normal mode analysis validated the binding affinity of the compounds and also revealed the strong stability and compactness of phytochemicals at the docked site. Additionally, DFT and DOS analyses were done to study the reactivity of the compounds and to further validate the selected phytochemicals. These results suggest that the identified phytochemicals possess high inhibitory potential against the target EGFR and can treat breast cancer. However, further in vitro and in vivo investigations are warranted towards the development of these constituents into novel anti-cancer drugs.Communicated by Ramaswamy H. Sarma.

A flexible data-free framework for structure-based de novo drug design with reinforcement learning

Contemporary structure-based molecular generative methods have demonstrated their potential to model the geometric and energetic complementarity between ligands and receptors, thereby facilitating the design of molecules with favorable binding affinity and target specificity. Despite the introduction of deep generative models for molecular generation, the atom-wise generation paradigm that partially contradicts chemical intuition limits the validity and synthetic accessibility of the generated molecules. Additionally, the dependence of deep learning models on large-scale structural data has hindered their adaptability across different targets. To overcome these challenges, we present a novel search-based framework, 3D-MCTS, for structure-based de novo drug design. Distinct from prevailing atom-centric methods, 3D-MCTS employs a fragment-based molecular editing strategy. The fragments decomposed from small-molecule drugs are recombined under predefined retrosynthetic rules, offering improved drug-likeness and synthesizability, overcoming the inherent limitations of atom-based approaches. Leveraging multi-threaded parallel simulations combined with a real-time energy constraint-based pruning strategy, 3D-MCTS achieves remarkable efficiency. At a fixed computational cost, it outperforms other state-of-the-art (SOTA) methods by producing molecules with enhanced binding affinity. Furthermore, its fragment-based approach ensures the generation of more dependable binding conformations, exhibiting a success rate 43.6% higher than that of other SOTAs. This advantage becomes even more pronounced when handling targets that significantly deviate from the training dataset. 3D-MCTS is capable of achieving thirty times more hits with high binding affinity than traditional virtual screening methods, which demonstrates the superior ability of 3D-MCTS to explore chemical space. Moreover, the flexibility of our framework makes it easy to incorporate domain knowledge during the process, thereby enabling the generation of molecules with desirable pharmacophores and enhanced binding affinity. The adaptability of 3D-MCTS is further showcased in metalloprotein applications, highlighting its potential across various drug design scenarios.

How Good Are Current Docking Programs at Nucleic Acid-Ligand Docking? A Comprehensive Evaluation

Nucleic acid (NA)-ligand interactions are of paramount importance in a variety of biological processes, including cellular reproduction and protein biosynthesis, and therefore, NAs have been broadly recognized as potential drug targets. Understanding NA-ligand interactions at the atomic scale is essential for investigating the molecular mechanism and further assisting in NA-targeted drug discovery. Molecular docking is one of the predominant computational approaches for predicting the interactions between NAs and small molecules. Despite the availability of versatile docking programs, their performance profiles for NA-ligand complexes have not been thoroughly characterized. In this study, we first compiled the largest structure-based NA-ligand binding data set to date, containing 800 noncovalent NA-ligand complexes with clearly identified ligands. Based on this extensive data set, eight frequently used docking programs, including six protein-ligand docking programs (LeDock, Surflex-Dock, UCSF Dock6, AutoDock, AutoDock Vina, and PLANTS) and two specific NA-ligand docking programs (rDock and RLDOCK), were systematically evaluated in terms of binding pose and binding affinity predictions. The results demonstrated that some protein-ligand docking programs, specifically PLANTS and LeDock, produced more promising or comparable results compared with the specialized NA-ligand docking programs. Among the programs evaluated, PLANTS, rDock, and LeDock showed the highest performance in binding pose prediction, and their top-1 and best root-mean-square deviation (rmsd) success rates were as follows: PLANTS (35.93 and 76.05%), rDock (27.25 and 72.16%), and LeDock (27.40 and 64.37%). Compared with the moderate level of binding pose prediction, few programs were successful in binding affinity prediction, and the best correlation (Rp = -0.461) was observed with PLANTS. Finally, further comparison with the latest NA-ligand docking program (NLDock) on four well-established data sets revealed that PLANTS and LeDock outperformed NLDock in terms of binding pose prediction on all data sets, demonstrating their significant potential for NA-ligand docking. To the best of our knowledge, this study is the most comprehensive evaluation of popular molecular docking programs for NA-ligand systems.