Efficient iterative virtual screening with Apache Spark and conformal prediction

Ahmed, Laeeq; Georgiev, Valentin; Capuccini, Marco; Toor, Salman; Schaal, Wesley; Laure, Erwin; Spjuth, Ola

doi:10.1186/s13321-018-0265-z

Cited by 37 publications

(36 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, what compounds should be picked at each iteration still remains a largely unresolved question. Hence, learning patterns in the screening data using artificial intelligence to increase hit rates, and hence discover active molecules faster and more efficiently, is gaining increasing attention [19][20][21][22] .…”

Section: Introductionmentioning

confidence: 99%

Discovering Highly Potent Molecules from an Initial Set of Inactives Using Iterative Screening

Cortés-Ciriano

Firth

Bender

et al. 2018

J. Chem. Inf. Model.

View full text Add to dashboard Cite

The versatility of similarity searching and quantitative structure-activity relationships to model the activity of compound sets within given bioactivity ranges (i.e., interpolation) is well established. However, their relative performance in the common scenario in early stage drug discovery where lots of inactive data but no active data points are available (i.e., extrapolation from the low-activity to the high-activity range) has not been thoroughly examined yet. To this aim, we have designed an iterative virtual screening strategy which was evaluated on 25 diverse bioactivity data sets from ChEMBL. We benchmark the efficiency of random forest (RF), multiple linear regression, ridge regression, similarity searching, and random selection of compounds to identify a highly active molecule in the test set among a large number of low-potency compounds. We use the number of iterations required to find this active molecule to evaluate the performance of each experimental setup. We show that linear and ridge regression often outperform RF and similarity searching, reducing the number of iterations to find an active compound by a factor of 2 or more. Even simple regression methods seem better able to extrapolate to high-bioactivity ranges than RF, which only provides output values in the range covered by the training set. In addition, examination of the scaffold diversity in the data sets used shows that in some cases similarity searching and RF require two times as many iterations as random selection depending on the chemical space covered in the initial training data. Lastly, we show using bioactivity data for COX-1 and COX-2 that our framework can be extended to multitarget drug discovery, where compounds are selected by concomitantly considering their activity against multiple targets. Overall, this study provides an approach for iterative screening where only inactive data are present in early stages of drug discovery in order to discover highly potent compounds and the best experimental set up in which to do so.

show abstract

Section: Introductionmentioning

confidence: 99%

Discovering Highly Potent Molecules from an Initial Set of Inactives Using Iterative Screening

Cortés-Ciriano

Firth

Bender

et al. 2018

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…Such new services (comprising models for new receptors) can be deployed in a similar way as shown for the reference instance on Kubernetes (code and instructions available on [41]). In the supplementary material we show how users can build models using our previous method [37] and then use the models to create service for a new receptor. Instructions are provided to deploy and add the Docker container for a new receptor to the service [39].…”

Section: Discussionmentioning

confidence: 99%

Predicting Target Profiles with Confidence as a Service using Docking Scores

Ahmed

Suardi

Arvidsson

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Background Identifying and assessing ligand-target binding is a core component in early drug discovery as one or more unwanted interactions may be associated with safety issues. Contributions We present an open-source, extendable web service for predicting target profiles with confidence using machine learning for a panel of 7 targets, where models are trained on molecular docking scores from a large virtual library. The method uses conformal prediction to produce valid measures of prediction efficiency for a particular confidence level. The service also offers the possibility to dock chemical structures to the panel of targets with QuickVina on individual compound basis. Results The docking procedure and resulting models were validated by docking well-known inhibitors for each of the 7 targets using QuickVina. The model predictions showed comparable performance to molecular docking scores against an external validation set. The implementation as publicly available microservices on Kubernetes ensures resilience, scalability, and extensibility.

show abstract

“…To confront this challenge, back in 2006 we introduced Progressive Docking -a hybrid docking/ machine learning approach utilizing 3D 'inductive' QSAR (quantitative structure-activity relationship) descriptors [8][9][10] to filter out molecules predicted to have unfavorable Glide docking scores 11 . While this method resulted in 3-4 fold enrichment of virtual hits (as several similar approaches reported more recently 12,13 ), the Progressive Docking did not gain a particular momentum. For that we could contemplate two possible reasons -on one hand, back in 2006 the ZINC database contained only ~1 million entries, which were amendable to conventional docking.…”

mentioning

confidence: 88%

Deep Docking - a Deep Learning Approach for Virtual Screening of Big Chemical Datasets

Gentile

Agrawal

Hsing

et al. 2019

Preprint

View full text Add to dashboard Cite

Drug discovery is an extensive and rigorous process that requires up to 2 billion dollars of investments and more than ten years of research and development to bring a molecule "from bench to a bedside". While virtual screening can significantly enhance drug discovery workflow, it ultimately lags the current rate of expansion of chemical databases that already incorporate billions of purchasable compounds. This surge of available small molecules presents great opportunities for drug discovery but also demands for faster virtual screening methods and protocols. In order to address this challenge, we herein introduce Deep Docking (D 2 ) -a novel deep learning-based approach which is suited for docking billions of molecular structures. The developed D 2 -platform utilizes quantitative structure-activity relationship (QSAR) based deep models trained on docking scores of subsets of a large chemical library (Big Base) to approximate the docking outcome for yet unprocessed molecular entries and to remove unfavorable structures in an iterative manner. We applied D 2 to virtually screen 1.36 billion molecules form the ZINC15 library against 12 prominent target proteins, and demonstrated up to 100-fold chemical data reduction and 6,000-fold enrichment for top hits, without notable loss of well-docked entities. The developed D 2 protocol can readily be used in conjunction with any docking program and was made publicly available.Drug discovery is an expensive and time-demanding process that faces many challenges, including low hit rates of high-throughput screening approaches among many others 1-2 . Methods of computer-aided drug discovery (CADD) can significantly speed up the pace of screening and help improving hit discovery rates 3 . Molecular docking is routinely used to process virtual libraries containing millions of molecular structures against a variety of drug targets with known three-

show abstract

Efficient iterative virtual screening with Apache Spark and conformal prediction

Cited by 37 publications

References 28 publications

Discovering Highly Potent Molecules from an Initial Set of Inactives Using Iterative Screening

Discovering Highly Potent Molecules from an Initial Set of Inactives Using Iterative Screening

Predicting Target Profiles with Confidence as a Service using Docking Scores

Deep Docking - a Deep Learning Approach for Virtual Screening of Big Chemical Datasets

Contact Info

Product

Resources

About