scholarly journals Deep Generative Design with 3D Pharmacophoric Constraints

2021 ◽  
Author(s):  
Fergus Imrie ◽  
Thomas E. Hadfield ◽  
Anthony R. Bradley ◽  
Charlotte M. Deane
Keyword(s):  
Design Process ◽  
Large Scale ◽  
Molecular Design ◽  
Structural Information ◽  
Three Dimensional ◽  
General Purpose ◽  
Generative Models ◽  
Base Line ◽  
Lead Optimisation ◽  
3D Representations

AbstractGenerative models have increasingly been proposed as a solution to the molecular design problem. However, it has proved challenging to control the design process or incorporate prior knowledge, limiting their practical use in drug discovery. In particular, generative methods have made limited use of three-dimensional (3D) structural information even though this is critical to binding. This work describes a method to incorporate such information and demonstrates the benefit of doing so. We combine an existing graph-based deep generative model, DeLinker, with a convolutional neural network to utilise physically-meaningful 3D representations of molecules and target pharmacophores. We apply our model, DEVELOP, to both linker and R-group design, demonstrating its suitability for both hit-to-lead and lead optimisation. The 3D pharmacophoric information results in improved generation and allows greater control of the design process. In multiple large-scale evaluations, we show that including 3D pharmacophoric constraints results in substantial improvements in the quality of generated molecules. On a challenging test set derived from PDBbind, our model improves the proportion of generated molecules with high 3D similarity to the original molecule by over 300%. In addition, DEVELOP recovers 10 × more of the original molecules compared to the base-line DeLinker method. Our approach is general-purpose, readily modifiable to alternate 3D representations, and can be incorporated into other generative frameworks. Code is available at https://github.com/oxpig/DEVELOP.

scholarly journals How to make sense of 3D representations for plant phenotyping: a compendium of processing and analysis techniques

2018 ◽  
Author(s):  
Breght Vandenberghe ◽  
Stephen Depuydt ◽  
Arnout Van Messem
Keyword(s):  
Machine Vision ◽  
Large Scale ◽  
Three Dimensional ◽  
Plant Phenotyping ◽  
Small Scale ◽  
Dimensional Approach ◽  
Crop Monitoring ◽  
Analysis Techniques ◽  
Further Development ◽  
3D Representations

Machine vision technology is moving more and more towards a three-dimensional approach, and plant phenotyping is following this trend. However, despite its potential, the complexity of the analysis of 3D representations has been the main bottleneck hindering the wider deployment of 3D plant phenotyping. In this review we provide an overview of typical steps for the processing and analysis of 3D representations of plants, to offer potential users of 3D phenotyping a first gateway into its application, and to stimulate its further development. We focus on plant phenotyping applications where the goal is to measure characteristics of single plants or crop canopies on a small scale in research settings, as opposed to large scale crop monitoring in the field.

Δ-Quantum machine learning for medicinal chemistry

2021 ◽  
Author(s):  
Kenneth Atz ◽  
Clemens Isert ◽  
Markus N. A. Böcker ◽  
José Jiménez-Luna ◽  
Gisbert Schneider
Keyword(s):  
Machine Learning ◽  
Message Passing ◽  
Density Functional ◽  
Large Scale ◽  
Molecular Design ◽  
Low Cost ◽  
Computational Cost ◽  
Three Dimensional ◽  
Biomolecular Systems ◽  
Quantum Observables

Many molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like molecules currently renders large-scale applications of quantum chemistry challenging. Aiming to mitigate this problem, we developed DelFTa, an open-source toolbox for the prediction of electronic properties of drug-like molecules at the density functional (DFT) level of theory, using Δ-machine-learning. Δ-Learning corrects the prediction error (Δ) of a fast but inaccurate property calculation. DelFTa employs state-of-the-art three-dimensional message-passing neural networks trained on a large dataset of QM properties. It provides access to a wide array of quantum observables on the molecular, atomic and bond levels by predicting approximations to DFT values from a low-cost semiempirical baseline. Δ-Learning outperformed its direct-learning counterpart for most of the considered QM endpoints. The results suggest that predictions for non-covalent intra- and intermolecular interactions can be extrapolated to larger biomolecular systems. The software is fully open-sourced and features documented command-line and Python APIs.

Quantification of Symmetry, Parallelism, and Continuity as Continuous Design Variables for Three-Dimensional Product Representations

2016 ◽  
Author(s):  
Ambrosio Valencia-Romero ◽  
José E. Lugo
Keyword(s):  
Design Process ◽  
Online Survey ◽  
Three Dimensional ◽  
Product Representation ◽  
Gestalt Principles ◽  
Logit Regression ◽  
Irregular Shapes ◽  
Design Variables ◽  
Future Work ◽  
3D Representations

This work introduces a methodology to quantify the form of a three-dimensional (3D) product representation using the Gestalt principles of symmetry, parallelism, and continuity, and how they can be used as descriptive parameters in product design. First, consistent quantifications of these three Gestalt principles for parametrized 3D representations in a zero-one scale are presented. Then, a generalized methodology applicable to any product form is discussed. It starts with the identification of important aesthetic forms of the product shape and the Gestalt principles that best related to those forms, and ends with the quantification of these Gestalt principles of a 3D product representation. The expressions to quantify the Gestalt principles in question are validated through an online survey in which subjects indicated how much they recognize symmetry, parallelism, or continuity from irregular shapes. Finally, random-effects ordered logit regression is used to determine if the expressions effectively describe the level of recognition of each Gestalt principle. Results show that the proposed quantifications for symmetry, parallelism, and continuity are congruent with subjects perception of these Gestalt principles, and the implications for designers and future work are discussed. Further implications in the design process of these quantifications include the optimization of the product shape for aesthetic, semantic, and functional goals.

scholarly journals Computer-aided Design of Peptidomimetic Inhibitors of falcipain-3: QSAR and Pharmacophore Models

2021 ◽  
Author(s):  
Boris D. Bekono ◽  
Akori Esmel ◽  
Brice Dali ◽  
Fidele Ntie-Kang ◽  
Melalie Keita ◽  
...  
Keyword(s):  
Computer Aided Design ◽  
Molecular Design ◽  
Structural Information ◽  
Three Dimensional ◽  
Pharmacophore Model ◽  
Combinatorial Libraries ◽  
3D Qsar ◽  
3D Models ◽  
Qsar Model ◽  
Computer Aided

In this work antiparasitic peptidomimetics inhibitors (PEP) of falcipain-3 (FP3) of Plasmodium falciparum (Pf) have been proposed using structure-based and computer-aided molecular design. Beginning with the crystal structure of PfFP3-K11017 complex (PDB ID: 3BWK), three-dimensional (3D) models of FP3-PEPx complexes with known activities (IC50exp) were prepared by in situ modification, based on molecular mechanics and implicit solvation to compute Gibbs free energies (GFE) of inhibitor-FP3 complex formation. This resulted in a quantitative structure-activity relationships (QSAR) model based on a linear correlation between computed GFE (ΔΔGcomp) and the experimentally measured IC50exp: (pIC50exp=-(IC50exp/109) =-0.4517×∆∆Gcomp+4.0865 ; R2 = 0.89). Apart from the structure-based relationship, a ligand-based quantitative pharmacophore model (PH4) of novel PEP analogs where substitutions were directed by comparative analysis of the active site interactions was derived using the proposed bound conformations of the PEPx. This provided structural information useful for the design of virtual combinatorial libraries (VL), which was virtually screened based on the 3D-QSAR PH4. The end results were predictory inhibitory activities falling within the low nanomolar concentration range.

Virtual Demonstration Tests of Large-Scale and Complex Artifacts Using an Open Source Parallel CAE System, ADVENTURE

2006 ◽  
Vol 110 ◽  
pp. 133-142 ◽  
Author(s):  
Shinobu Yoshimura
Keyword(s):  
Open Source ◽  
Large Scale ◽  
Three Dimensional ◽  
Element Model ◽  
General Purpose ◽  
Full Scale ◽  
Grid Environments ◽  
Dimensional Finite Element ◽  
Three Dimensional Finite Element ◽  
Cae System

The ADVENTURE project started as one of the research projects in the "Computational Science & Engineering" field selected for the "Research for the Future" Program sponsored by the Japan Society for the Promotion of Science during 1997-2002. Since March 2002, the project has continued as an independent project. In the project we have been developing an advanced general-purpose computational mechanics system, named ADVENTURE, running in various kinds of parallel and ditributed environments. The system is designed to be able to analyze a three-dimensional finite element model of arbitrary shape with 10-100 million DOFs mesh, and additionally to enable parametric and non-parametric shape optimization. The first version of the system has been released from the project website as open source software since March, 2002. 2,049 registered users in academia and industries have downloaded 12,827 modules and been using them, while one company has developed and released its commercial version named ADVENTUREcluster. The ADVENTURE system has been successfully implemented in various types of parallel and distributed environments including PC clusters, massively parallel processers such as Hitachi SR8000/MPP and the Earth Simulator, and Grid environments such as ITBL (IT-based Laboratory). The system has been successfully applied to solve various real world problems such as response of a full scale nuclear pressure vessel model and thermoelastic deformation of full scale electric mounting board of a mobile PC.

Digital soil: Robotically 3D-printed granular bio-composites

2020 ◽  
Vol 18 (2) ◽  
pp. 194-211
Author(s):  
Daniela Mitterberger ◽  
Tiziano Derme
Keyword(s):  
Design Process ◽  
Manufacturing Process ◽  
Large Scale ◽  
Soil Aggregates ◽  
Three Dimensional ◽  
Digital Fabrication ◽  
Material Research ◽  
Geometrical Properties ◽  
Robotic Fabrication ◽  
Material Systems

Organic granular materials offer a valid alternative for non-biodegradable composites widely adopted in building construction and digital fabrication. Despite the need to find alternatives to fuel-based solutions, current material research in architecture mostly supports strategies that favour predictable, durable and homogeneous solutions. Materials such as soil, due to their physical properties and volatile nature, present new challenges and potentials to change the way we manufacture, built and integrate material systems and environmental factors into the design process. This article proposes a novel fabrication framework that combines high-resolution three-dimensional-printed biodegradable materials with a novel robotic-additive manufacturing process for soil structures. Furthermore, the research reflects on concepts such as affordance and tolerance within the field of digital fabrication, especially in regards to bio-materials and robotic fabrication. Soil as a building material has a long tradition. New developments in earth construction show how earthen buildings can create novel, adaptive and sustainable structures. Nevertheless, existing large-scale earthen construction methods can only produce highly simplified shapes with rough geometrical articulations. This research proposes to use a robotic binder-jetting process that creates novel organic bio-composites to overcome such limitations of common earth constructions. In addition, this article shows how biological polymers, such as polysaccharides-based hydrogels, can be used as sustainable, biodegradable binding agents for soil aggregates. This article is divided into four main sections: architecture and affordance; tolerance versus precision; water-based binders; and robotic fabrication parameters. Digital Soil envisions a shift in the design practice and digital fabrication that builds on methods for tolerance handling. In this context, material and geometrical properties such as material porosity, hydraulic conductivity and natural evaporation rate affect the architectural resolution, introducing a design process driven by matter. Digital Soil shows the potential of a fully reversible biodegradable manufacturing process for load-bearing architectural elements, opening up new fields of application for sustainable material systems that can enhance the ecological potential of architectural construction.

scholarly journals Computer-Aided Design of Peptidomimetic Inhibitors of Falcipain-3: QSAR and Pharmacophore Models

2021 ◽  
Vol 89 (4) ◽  
pp. 44
Author(s):  
Boris D. Bekono ◽  
Akori E. Esmel ◽  
Brice Dali ◽  
Fidele Ntie-Kang ◽  
Melalie Keita ◽  
...  
Keyword(s):  
Computer Aided Design ◽  
Molecular Design ◽  
Structural Information ◽  
Three Dimensional ◽  
Pharmacophore Model ◽  
Combinatorial Libraries ◽  
3D Qsar ◽  
3D Models ◽  
Qsar Model ◽  
Computer Aided

In this work, antiparasitic peptidomimetics inhibitors (PEP) of falcipain-3 (FP3) of Plasmodium falciparum (Pf) are proposed using structure-based and computer-aided molecular design. Beginning with the crystal structure of PfFP3-K11017 complex (PDB ID: 3BWK), three-dimensional (3D) models of FP3-PEPx complexes with known activities () were prepared by in situ modification, based on molecular mechanics and implicit solvation to compute Gibbs free energies (GFE) of inhibitor-FP3 complex formation. This resulted in a quantitative structure–activity relationships (QSAR) model based on a linear correlation between computed GFE () and the experimentally measured . Apart from the structure-based relationship, a ligand-based quantitative pharmacophore model (PH4) of novel PEP analogues where substitutions were directed by comparative analysis of the active site interactions was derived using the proposed bound conformations of the PEPx. This provided structural information useful for the design of virtual combinatorial libraries (VL), which was virtually screened based on the 3D-QSAR PH4. The end results were predictive inhibitory activities falling within the low nanomolar concentration range.

scholarly journals Deep Generative Models for 3D Compound Design

2019 ◽  
Author(s):  
Fergus Imrie ◽  
Anthony R. Bradley ◽  
Mihaela van der Schaar ◽  
Charlotte M. Deane
Keyword(s):  
Large Scale ◽  
Structural Information ◽  
Generative Models ◽  
Generative Model ◽  
Machine Learning Techniques ◽  
Generation Process ◽  
Diverse Range ◽  
Partial Structures ◽  
Compound Design ◽  
3D Information

AbstractRational compound design remains a challenging problem for both computational methods and medicinal chemists. Computational generative methods have begun to show promising results for the design problem. However, they have not yet used the power of 3D structural information. We have developed a novel graph-based deep generative model that combines state-of-the-art machine learning techniques with structural knowledge. Our method (“DeLinker”) takes two fragments or partial structures and designs a molecule incorporating both. The generation process is protein context dependent, utilising the relative distance and orientation between the partial structures. This 3D information is vital to successful compound design, and we demonstrate its impact on the generation process and the limitations of omitting such information. In a large scale evaluation, DeLinker designed 60% more molecules with high 3D similarity to the original molecule than a database baseline. When considering the more relevant problem of longer linkers with at least five atoms, the outperformance increased to 200%. We demonstrate the effectiveness and applicability of this approach on a diverse range of design problems: fragment linking, scaffold hopping, and proteolysis targeting chimera (PROTAC) design. As far as we are aware, this is the first molecular generative model to incorporate 3D structural information directly in the design process. Code is available at https://github.com/oxpig/DeLinker.

Rockfall on the Slope of the Embankment of Finite Element Analysis of the Impact of the Impact

2011 ◽  
Vol 261-263 ◽  
pp. 1119-1123
Author(s):  
Zi Jian Wang ◽  
Sheng Xie Xiao
Keyword(s):  
Finite Element ◽  
Large Scale ◽  
Three Dimensional ◽  
Road Construction ◽  
General Purpose ◽  
Simulation Software ◽  
Initial State ◽  
Element Analysis ◽  
The Road ◽  
The Impact

This article in a typical road slope model as an example, through large-scale general-purpose finite element simulation software ANSYS/LS-DYNA rockfall on the slope of the roadbed impact for three-dimensional dynamic analysis in the roadbed, slope geological parameters and falling stone initial state is not at the same time on the roadbed impact the results of comparative analysis. The result is the mountain slope of the road construction project provides reference.

Large-Scale Analysis and Local Stress Assessment of Flexible Unbonded Pipes Using FEA

2012 ◽  
Author(s):  
Ben Edmans ◽  
Giulio Alfano ◽  
Hamid Bahai
Keyword(s):  
Large Scale ◽  
Three Dimensional ◽  
Local Stress ◽  
General Purpose ◽  
Operating Conditions ◽  
Global Model ◽  
Scale Analysis ◽  
Analysis Model ◽  
Detailed Model ◽  
Stress Prediction

Lifespan assessment of flexible risers involves the evaluation of fatigue parameters, requiring accurate predictions of stresses and their variation in pipe components. For predicting the effect of complex three-dimensional nonlinear dynamics on component stress histories, multi-scale methods can combine generality of application with computational efficiency. In this paper, we describe the development of a two-scale computational homogenisation procedure linking a coarse-scale analysis model using specialised beam elements, and a detailed stress prediction model consisting of a pipe section with components modelled with shell elements and frictional contact interactions. To use the procedure, the detailed model first functions as a virtual test rig, by which parameters of the global model may be determined. For detailed stress prediction, the global model is tested under the operating conditions of interest providing a set of generalised strains which are applied to the detailed model. The models are implemented in the general-purpose finite-element package Abaqus. As key aspects of the procedure, we show how generalised stresses and strains can be imposed on the detailed model uniformly without introducing spurious boundary effects and demonstrate how local stresses can be determined using strain data from the global model.

Sign in / Sign up

Export Citation Format

Share Document