De Novo Molecular Design of Caspase-6 Inhibitors by a GRU-Based Recurrent Neural Network Combined with a Transfer Learning Approach.
Ontology highlight
ABSTRACT:
SUBMITTER: Huang S
PROVIDER: S-EPMC8706867 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:The global pandemic of coronavirus disease (COVID-19) caused by SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) created a rush to discover drug candidates. Despite the efforts, so far no vaccine or drug has been approved for treatment. Artificial intelligence offers solutions that could accelerate the discovery and optimization of new antivirals, especially in the current scenario dominated by the scarcity of compounds active against SARS-CoV-2. The main protease (Mpro) of SARS-CoV-2 is an attractive target for drug discovery due to the absence in humans and the essential role in viral replication. In this work, we developed a deep learning platform for de novo design of putative inhibitors of SARS-CoV-2 main protease (Mpro). Our methodology consists of 3 main steps: (1) training and validation of general chemistry-based generative model; (2) fine-tuning of the generative model for the chemical space of SARS-CoV- Mpro inhibitors and (3) training of a classifier for bioactivity prediction using transfer learning. The fine-tuned chemical model generated > 90% valid, diverse and novel (not present on the training set) structures. The generated molecules showed a good overlap with Mpro chemical space, displaying similar physicochemical properties and chemical structures. In addition, novel scaffolds were also generated, showing the potential to explore new chemical series. The classification model outperformed the baseline area under the precision-recall curve, showing it can be used for prediction. In addition, the model also outperformed the freely available model Chemprop on an external test set of fragments screened against SARS-CoV-2 Mpro, showing its potential to identify putative antivirals to tackle the COVID-19 pandemic. Finally, among the top-20 predicted hits, we identified nine hits via molecular docking displaying binding poses and interactions similar to experimentally validated inhibitors.
| S-EPMC7852053 | biostudies-literature
Project description:A recurrent neural network (RNN) is a machine learning model that learns the relationship between elements of an input series, in addition to inferring a relationship between the data input to the model and target output. Memory augmentation allows the RNN to learn the interrelationships between elements of the input over a protracted length of the input series. Inspired by the success of stack augmented RNN (StackRNN) to generate strings for various applications, we present two memory augmented RNN-based architectures: the Neural Turing Machine (NTM) and the Differentiable Neural Computer (DNC) for the de-novo generation of small molecules. We trained a character-level convolutional neural network (CNN) to predict the properties of a generated string and compute a reward or loss in a deep reinforcement learning setup to bias the Generator to produce molecules with the desired property. Further, we compare the performance of these architectures to gain insight to their relative merits in terms of the validity and novelty of the generated molecules and the degree of property bias towards the computational generation of de-novo drugs. We also compare the performance of these architectures with simpler recurrent neural networks (Vanilla RNN, LSTM, and GRU) without an external memory component to explore the impact of augmented memory in the task of de-novo generation of small molecules.
| S-EPMC9223405 | biostudies-literature
Project description:Research productivity in the pharmaceutical industry has declined significantly in recent decades, with higher costs, longer timelines, and lower success rates of drug candidates in clinical trials. This has prioritized the scalability and multiobjectivity of drug discovery and design. De novo drug design has emerged as a promising approach; molecules are generated from scratch, thus reducing the reliance on trial and error and premade molecular repositories. However, optimizing for molecular traits remains challenging, impeding the implementation of de novo methods. In this work, we propose a de novo approach capable of optimizing multiple traits collectively. A recurrent neural network was used to generate molecules which were then ranked based on multiple properties by a nondominated sorting algorithm. The best of the molecules generated were selected and used to fine-tune the recurrent neural network through transfer learning, creating a cycle that mimics the traditional design–synthesis–test cycle. We demonstrate the efficacy of this approach through a proof of concept, optimizing for constraints on molecular weight, octanol-water partition coefficient, the number of rotatable bonds, hydrogen bond donors, and hydrogen bond acceptors simultaneously. Analysis of the molecules generated after five iterations of the cycle revealed a 14-fold improvement in the quality of generated molecules, along with improvements to the accuracy of the recurrent neural network and the structural diversity of the molecules generated. This cycle notably does not require large amounts of training data nor any handwritten scoring functions. Altogether, this approach uniquely combines scalable generation with multiobjective optimization of molecules.
| S-EPMC7026957 | biostudies-literature
Project description:Bilins are linear tetrapyrrole chromophores with a wide range of visible and near-visible light absorption and emission properties. These properties are tuned upon binding to natural proteins and exploited in photosynthetic light-harvesting and non-photosynthetic light-sensitive signalling. These pigmented proteins are now being manipulated to develop fluorescent experimental tools. To engineer the optical properties of bound bilins for specific applications more flexibly, we have used first principles of protein folding to design novel, stable and highly adaptable bilin-binding four-?-helix bundle protein frames, called maquettes, and explored the minimal requirements underlying covalent bilin ligation and conformational restriction responsible for the strong and variable absorption, fluorescence and excitation energy transfer of these proteins. Biliverdin, phycocyanobilin and phycoerythrobilin bind covalently to maquette Cys in vitro A blue-shifted tripyrrole formed from maquette-bound phycocyanobilin displays a quantum yield of 26%. Although unrelated in fold and sequence to natural phycobiliproteins, bilin lyases nevertheless interact with maquettes during co-expression in Escherichia coli to improve the efficiency of bilin binding and influence bilin structure. Bilins bind in vitro and in vivo to Cys residues placed in loops, towards the amino end or in the middle of helices but bind poorly at the carboxyl end of helices. Bilin-binding efficiency and fluorescence yield are improved by Arg and Asp residues adjacent to the ligating Cys on the same helix and by His residues on adjacent helices.
| S-EPMC5938588 | biostudies-literature
Project description:The COVID-19 pandemic prompted many scientists to investigate remedies against SARS-CoV-2 and related viruses that are likely to appear in the future. As the main protease of the virus, MPro, is highly conserved among coronaviruses, it has emerged as a prime target for developing inhibitors. Using a combination of virtual screening and molecular modeling, we identified small molecules that were easily accessible and could be quickly diversified. Biochemical assays confirmed a class of pyridones as low micromolar non-covalent inhibitors of the viral main protease.
| S-EPMC8916680 | biostudies-literature
Project description:Histones are small proteins critical to the efficient packaging of DNA in the nucleus. DNA–protein complexes, known as nucleosomes, are formed when the DNA winds itself around the surface of the histones. The methylation of histone residues by enhancer of zeste homolog 2 (EZH2) maintains gene repression over successive cell generations. Overexpression of EZH2 can silence important tumor suppressor genes leading to increased invasiveness of many types of cancers. This makes the inhibition of EZH2 an important target in the development of cancer therapeutics. We employed a three-stage computational de novo peptide design method to design inhibitory peptides of EZH2. The method consists of a sequence selection stage and two validation stages for fold specificity and approximate binding affinity. The sequence selection stage consists of an integer linear optimization model that was solved to produce a rank-ordered list of amino acid sequences with increased stability in the bound peptide-EZH2 structure. These sequences were validated through the calculation of the fold specificity and approximate binding affinity of the designed peptides. Here we report the discovery of novel EZH2 inhibitory peptides using the de novo peptide design method. The computationally discovered peptides were experimentally validated in vitro using dose titrations and mechanism of action enzymatic assays. The peptide with the highest in vitro response, SQ037, was validated in nucleo using quantitative mass spectrometry-based proteomics. This peptide had an IC50 of 13.5 mM, demonstrated greater potency as an inhibitor when compared to the native and K27A mutant control peptides, and demonstrated competitive inhibition versus the peptide substrate. Additionally, this peptide demonstrated high specificity to the EZH2 target in comparison to other histone methyltransferases. The validated peptides are the first computationally designed peptides that directly inhibit EZH2. These inhibitors should prove useful for further chromatin biology investigations.
| S-EPMC3989331 | biostudies-literature
Project description:Histones are small proteins critical to the efficient packaging of DNA in the nucleus. DNA-protein complexes, known as nucleosomes, are formed when the DNA winds itself around the surface of the histones. The methylation of histone residues by enhancer of zeste homolog 2 (EZH2) maintains gene repression over successive cell generations. Overexpression of EZH2 can silence important tumor suppressor genes leading to increased invasiveness of many types of cancers. This makes the inhibition of EZH2 an important target in the development of cancer therapeutics. We employed a three-stage computational de novo peptide design method to design inhibitory peptides of EZH2. The method consists of a sequence selection stage and two validation stages for fold specificity and approximate binding affinity. The sequence selection stage consists of an integer linear optimization model that was solved to produce a rank-ordered list of amino acid sequences with increased stability in the bound peptide-EZH2 structure. These sequences were validated through the calculation of the fold specificity and approximate binding affinity of the designed peptides. Here we report the discovery of novel EZH2 inhibitory peptides using the de novo peptide design method. The computationally discovered peptides were experimentally validated in vitro using dose titrations and mechanism of action enzymatic assays. The peptide with the highest in vitro response, SQ037, was validated in nucleo using quantitative mass spectrometry-based proteomics. This peptide had an IC50 of 13.5 [Formula: see text]M, demonstrated greater potency as an inhibitor when compared to the native and K27A mutant control peptides, and demonstrated competitive inhibition versus the peptide substrate. Additionally, this peptide demonstrated high specificity to the EZH2 target in comparison to other histone methyltransferases. The validated peptides are the first computationally designed peptides that directly inhibit EZH2. These inhibitors should prove useful for further chromatin biology investigations.
| S-EPMC3938834 | biostudies-literature
Project description:The impact of the ravages of COVID-19 on people's lives is obvious, and the development of novel potential inhibitors against SARS-CoV-2 main protease (Mpro), which has been validated as a potential target for drug design, is urgently needed. This study developed a model named MproI-GEN, which can be used for the de novo design of potential Mpro inhibitors (MproIs) based on deep learning. The model was mainly composed of long-short term memory modules, and the last layer was re-trained with transfer learning. The validity (0.9248), novelty (0.9668), and uniqueness (0.0652) of the designed potential MproI library (PMproIL) were evaluated, and the results showed that MproI-GEN could be used to design structurally novel and reasonable molecules. Additionally, PMproIL was filtered based on machine learning models and molecular docking. After filtering, the potential MproIs were verified with molecular dynamics simulations to evaluate the binding stability levels of these MproIs and SARS-CoV-2 Mpro, thereby illustrating the inhibitory effects of the potential MproIs against Mpro. Two potential MproIs were proposed in this study. This study provides not only new possibilities for the development of COVID-19 drugs but also a complete pipeline for the discovery of novel lead compounds.
| S-EPMC9197785 | biostudies-literature
Project description:Natural metalloproteins perform many functions - ranging from sensing to electron transfer and catalysis - in which the position and property of each ligand and metal, is dictated by protein structure. De novo protein design aims to define an amino acid sequence that encodes a specific structure and function, providing a critical test of the hypothetical inner workings of (metallo)proteins. To date, de novo metalloproteins have used simple, symmetric tertiary structures - uncomplicated by the large size and evolutionary marks of natural proteins - to interrogate structure-function hypotheses. In this Review, we discuss de novo design applications, such as proteins that induce complex, increasingly asymmetric ligand geometries to achieve function, as well as the use of more canonical ligand geometries to achieve stability. De novo design has been used to explore how proteins fine-tune redox potentials and catalyse both oxidative and hydrolytic reactions. With an increased understanding of structure-function relationships, functional proteins including O2-dependent oxidases, fast hydrolases, and multi-proton/multi-electron reductases, have been created. In addition, proteins can now be designed using xeno-biological metals or cofactors and principles from inorganic chemistry to derive new-to-nature functions. These results and the advances in computational protein design suggest a bright future for the de novo design of diverse, functional metalloproteins.
| S-EPMC9264687 | biostudies-literature
Project description:Recent years have seen tremendous success in the design of novel drug molecules through deep generative models. Nevertheless, existing methods only generate drug-like molecules, which require additional structural optimization to be developed into actual drugs. In this study, a deep learning method for generating target-specific ligands was proposed. This method is useful when the dataset for target-specific ligands is limited. Deep learning methods can extract and learn features (representations) in a data-driven way with little or no human participation. Generative pretraining (GPT) was used to extract the contextual features of the molecule. Three different protein-encoding methods were used to extract the physicochemical properties and amino acid information of the target protein. Protein-encoding and molecular sequence information are combined to guide molecule generation. Transfer learning was used to fine-tune the pretrained model to generate molecules with better binding ability to the target protein. The model was validated using three different targets. The docking results show that our model is capable of generating new molecules with higher docking scores for the target proteins.
| S-EPMC9865828 | biostudies-literature