BRADSHAW: a system for automated molecular design

Green, Darren V. S.; Pickett, Stephen D.; Luscombe, Christopher N.; Senger, Stefan; Marcus, David J.; Meslamani, Jamel; Brett, David; Powell, Adam J.; Masson, Jonathan

doi:10.1007/s10822-019-00234-8

Cited by 51 publications

(53 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…MLPs for accelerated simulations and ML models that predict the outcome of QM calculations, since those are the most mature and offer a reasonable balance of usability and pay-off for industrial applications. More general applications of ML approaches, for example for retrosynthesis [69,70], direct prediction of experimental properties [71][72][73], and molecule generation and optimization [74][75][76], are discussed in references [77][78][79][80]. In the following section, we discuss different types of MLPs and approaches for their construction.…”

Section: Overcoming the Limitations Of Qm-based Simulations With Machmentioning

confidence: 99%

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Morawietz

Artrith

2020

J Comput Aided Mol Des

View full text Add to dashboard Cite

Atomistic simulations have become an invaluable tool for industrial applications ranging from the optimization of protein-ligand interactions for drug discovery to the design of new materials for energy applications. Here we review recent advances in the use of machine learning (ML) methods for accelerated simulations based on a quantum mechanical (QM) description of the system. We show how recent progress in ML methods has dramatically extended the applicability range of conventional QM-based simulations, allowing to calculate industrially relevant properties with enhanced accuracy, at reduced computational cost, and for length and time scales that would have otherwise not been accessible. We illustrate the benefits of ML-accelerated atomistic simulations for industrial R&D processes by showcasing relevant applications from two very different areas, drug discovery (pharmaceuticals) and energy materials. Writing from the perspective of both a molecular and a materials modeling scientist, this review aims to provide a unified picture of the impact of ML-accelerated atomistic simulations on the pharmaceutical, chemical, and materials industries and gives an outlook on the exciting opportunities that could emerge in the future.

show abstract

Section: Overcoming the Limitations Of Qm-based Simulations With Machmentioning

confidence: 99%

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Morawietz

Artrith

2020

J Comput Aided Mol Des

View full text Add to dashboard Cite

show abstract

“…The two most common methods for filtering molecules in de novo design are to apply a filter persistently through each iteration of an experiment, such that every scored molecule has necessarily passed all filters (Yuan et al, 2011;Green et al, 2019), or, to counter-screen at the end of the design process by filtering the final scored molecules (especially to augment linked generator-discriminators, as in Zhavoronkov et al, 2019). Interestingly, the modularity built into Deriver permits a third option: a delayed filtering mechanism, in which the algorithm is allowed to explore chemical space without filtering, until it is turned on by reaching some important threshold (typically a first convergence).…”

Section: (Figure 1: All Benchmarks)mentioning

confidence: 99%

“…It may also require training new generators for each discriminative task (Altae-Tran et al, 2017), and suffers from the black box nature of neural network predictions. To maximize utility and the ease in which a human expert can usefully participate in an AI guided design-make-test cycle (Green et al, 2019), it is critical that design pipelines be flexible, tunable, modular, and interpretable.…”

Section: Introductionmentioning

confidence: 99%

“…(2) the assessment of candidate molecules, including determining to what degree (if any) a candidate satisfies the constraints of the objective(s) ( discrimination ); and (3) the search strategy, which integrates generation and discrimination to guide chemical space traversal ( exploration ). While several studies (Polishchuk, 2020;Hartenfeller & Schneider, 2011;Green et al, 2019;Segler et al, 2017) have made similar distinctions between components of a de novo drug design algorithm, no system adequately leverages the separation of these concerns in their implementation, specifically with respect to decoupling the chemical space search (exploration).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Assessing Methods and Obstacles in Chemical Space Exploration

Reeves¹,

DiFrancesco²,

Shahani³

et al. 2020

Preprint

View full text Add to dashboard Cite

Benchmarking the performance of generative methods for drug design is complex and multifaceted. In this report, we propose a separation of concerns for de novo drug design, categorizing the task into three main categories: generation , discrimination , and exploration. We demonstrate that changes to any of these three concerns impacts benchmark performance for drug design tasks. In this report we present Deriver, an open-source Python package that acts as a modular framework for molecule generation, with a focus on integrating multiple generative methods. Using Deriver, we demonstrate that changing parameters related to each of these three concerns impacts chemical space traversal significantly, and that the freedom to independently adjust each is critical to real-world applications having conflicting priorities. We find that combining multiple generative methods can improve optimization of molecular properties and lower the chance of becoming trapped in local minima. Additionally, filtering molecules for drug-likeness (based on physicochemical properties and SMARTS pattern matching) before they are scored may hinder exploration, but can also improve the quality of the final molecules. Finally, we demonstrate that any given task has an exploration algorithm best suited to it, though in practice linear probabilistic sampling generally results in the best outcomes, when compared to Monte Carlo sampling or greedy sampling. Deriver is being made freely available, to help others interested in collaboratively improving existing methods in de novo drug design centered around inheritance of molecular structure, modularity, extensibility, and separation of concerns.

show abstract

“…In the pharmaceutical industry there are active efforts to use open data and to explore data driven computational platforms. [45,46,47,48] These efforts are not without difficulties, and are often attempted by single organizations, hence, may lack the inter-organizational standards needed completely open data. However, some benefits are being seen in this area in terms of insights from calculations and high throughput screening of lead molecule structures.…”

Section: Obtaining Reliable Physical Modelsmentioning

confidence: 99%

What Can Digitization Do For Formulated Product Innovation and Development

McDonagh¹,

Swope²,

Anderson³

et al. 2020

Preprint

View full text Add to dashboard Cite

Digitization oﬀers signiﬁcant opportunities for the formulated product industry to transform the way it works and develop new methods of business. R&D is one area of operation that is challenging to take advantage of these technologies due to its high level of domain specialisation and creativity but the beneﬁts could be signiﬁcant. Recent developments of base level technologies such as artiﬁcial intelligence (AI)/machine learning (ML), robotics and high performance computing (HPC), to name a few, present disruptive and transformative technologies which could oﬀer new insights, discovery methods and enhanced chemical control when combined in a digital ecosystem of connectivity, distributive services and decentralisation. At the fundamental level, research in these technologies has shown that new physical and chemical insights can be gained, which in turn can augment experimental R&D approaches through physics-based chemical simulation, data driven models and hybrid approaches. In all of these cases, high quality data is required to build and validate models in addition to the skills and expertise to exploit such methods. In this article we give an overview of some of the digital technology demonstrators we have developed for formulated product R&D. We discuss the challenges in building and deploying these demonstrators.<br>

show abstract

BRADSHAW: a system for automated molecular design

Cited by 51 publications

References 66 publications

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Assessing Methods and Obstacles in Chemical Space Exploration

What Can Digitization Do For Formulated Product Innovation and Development

Contact Info

Product

Resources

About