Search

Abstract Background The Blue Obelisk movement was established in 2005 as a response to the lack of Open Data, Open Standards and Open Source (ODOSOS) in chemistry. It aims to make it easier to carry out chemistry research by promoting interoperability between chemistry software, encouraging cooperation between Open Source developers, and developing community resources and Open Standards. Results This contribution looks back on the work carried out by the Blue Obelisk in the past 5 years and surveys progress and remaining challenges in the areas of Open Data, Open Standards, and Open Source in chemistry. Conclusions We show that the Blue Obelisk has been very successful in bringing together researchers and developers with common interests in ODOSOS, leading to development of many useful resources freely available to the chemistry community.

Abstract Background QSAR is a widely used method to relate chemical structures to responses or properties based on experimental observations. Much effort has been made to evaluate and validate the statistical modeling in QSAR, but these analyses treat the dataset as fixed. An overlooked but highly important issue is the validation of the setup of the dataset, which comprises addition of chemical structures as well as selection of descriptors and software implementations prior to calculations. This process is hampered by the lack of standards and exchange formats in the field, making it virtually impossible to reproduce and validate analyses and drastically constrain collaborations and re-use of data. Results We present a step towards standardizing QSAR analyses by defining interoperable and reproducible QSAR datasets, consisting of an open XML format (QSAR-ML) which builds on an open and extensible descriptor ontology. The ontology provides an extensible way of uniquely defining descriptors for use in QSAR experiments, and the exchange format supports multiple versioned implementations of these descriptors. Hence, a dataset described by QSAR-ML makes its setup completely reproducible. We also provide a reference implementation as a set of plugins for Bioclipse which simplifies setup of QSAR datasets, and allows for exporting in QSAR-ML as well as old-fashioned CSV formats. The implementation facilitates addition of new descriptor implementations from locally installed software and remote Web services; the latter is demonstrated with REST and XMPP Web services. Conclusions Standardized QSAR datasets open up new ways to store, query, and exchange data for subsequent analyses. QSAR-ML supports completely reproducible creation of datasets, solving the problems of defining which software components were used and their versions, and the descriptor ontology eliminates confusions regarding descriptors by defining them crisply. This makes is easy to join, extend, combine datasets and hence work collectively, but also allows for analyzing the effect descriptors have on the statistical model's performance. The presented Bioclipse plugins equip scientists with graphical tools that make QSAR-ML easily accessible for the community.

Conformal prediction has seen many applications in pharmaceutical science, being able to calibrate outputs of machine learning models and producing valid prediction intervals. We here present the open source software CPSign that is a complete implementation of conformal prediction for cheminformatics modeling. CPSign implements inductive and transductive conformal prediction for classification and regression, and probabilistic prediction with the Venn-ABERS methodology. The main chemical representation is signatures but other types of descriptors are also supported. The main modeling methodology is support vector machines (SVMs), but additional modeling methods are supported via an extension mechanism, e.g. DeepLearning4J models. We also describe features for visualizing results from conformal models including calibration and efficiency plots, as well as features to publish predictive models as REST services. We compare CPSign against other common cheminformatics modeling approaches including random forest, and a directed message-passing neural network. The results show that CPSign produces robust predictive performance with comparative predictive efficiency, with superior runtime and lower hardware requirements compared to neural network based models. CPSign has been used in several studies and is in production-use in multiple organizations. The ability to work directly with chemical input files, perform descriptor calculation and modeling with SVM in the conformal prediction framework, with a single software package having a low footprint and fast execution time makes CPSign a convenient and yet flexible package for training, deploying, and predicting on chemical data. CPSign can be downloaded from GitHub at https://github.com/arosbio/cpsign .Scientific contribution CPSign provides a single software that allows users to perform data preprocessing, modeling and make predictions directly on chemical structures, using conformal and probabilistic prediction. Building and evaluating new models can be achieved at a high abstraction level, without sacrificing flexibility and predictive performance-showcased with a method evaluation against contemporary modeling approaches, where CPSign performs on par with a state-of-the-art deep learning based model., (© 2024. The Author(s).)

The applicability domain of machine learning models trained on structural fingerprints for the prediction of biological endpoints is often limited by the lack of diversity of chemical space of the training data. In this work, we developed similarity-based merger models which combined the outputs of individual models trained on cell morphology (based on Cell Painting) and chemical structure (based on chemical fingerprints) and the structural and morphological similarities of the compounds in the test dataset to compounds in the training dataset. We applied these similarity-based merger models using logistic regression models on the predictions and similarities as features and predicted assay hit calls of 177 assays from ChEMBL, PubChem and the Broad Institute (where the required Cell Painting annotations were available). We found that the similarity-based merger models outperformed other models with an additional 20% assays (79 out of 177 assays) with an AUC > 0.70 compared with 65 out of 177 assays using structural models and 50 out of 177 assays using Cell Painting models. Our results demonstrated that similarity-based merger models combining structure and cell morphology models can more accurately predict a wide range of biological assay outcomes and further expanded the applicability domain by better extrapolating to new structural and morphology spaces., (© 2023. The Author(s).)

Polypharmacy refers to the administration of multiple drugs on a daily basis. It has demonstrated effectiveness in treating many complex diseases , but it has a higher risk of adverse drug reactions. Hence, the prediction of polypharmacy side effects is an essential step in drug testing, especially for new drugs. This paper shows that the current knowledge graph (KG) based state-of-the-art approach to polypharmacy side effect prediction does not work well for new drugs, as they have a low number of known connections in the KG. We propose a new method , SimVec, that solves this problem by enhancing the KG structure with a structure-aware node initialization and weighted drug similarity edges. We also devise a new 3-step learning process, which iteratively updates node embeddings related to side effects edges, similarity edges, and drugs with limited knowledge. Our model significantly outperforms existing KG-based models. Additionally, we examine the problem of negative relations generation and show that the cache-based approach works best for polypharmacy tasks., (© 2022. The Author(s).)

Comparing chemical structures to infer protein targets and functions is a common approach, but basing comparisons on chemical similarity alone can be misleading. Here we present a methodology for predicting target protein clusters using deep neural networks. The model is trained on clusters of compounds based on similarities calculated from combined compound-protein and protein-protein interaction data using a network topology approach. We compare several deep learning architectures including both convolutional and recurrent neural networks. The best performing method, the recurrent neural network architecture MolPMoFiT, achieved an F1 score approaching 0.9 on a held-out test set of 8907 compounds. In addition, in-depth analysis on a set of eleven well-studied chemical compounds with known functions showed that predictions were justifiable for all but one of the chemicals. Four of the compounds, similar in their molecular structure but with dissimilarities in their function, revealed advantages of our method compared to using chemical similarity., (© 2022. The Author(s).)

Confidence predictors can deliver predictions with the associated confidence required for decision making and can play an important role in drug discovery and toxicity predictions. In this work we investigate a recently introduced version of conformal prediction, synergy conformal prediction, focusing on the predictive performance when applied to bioactivity data. We compare the performance to other variants of conformal predictors for multiple partitioned datasets and demonstrate the utility of synergy conformal predictors for federated learning where data cannot be pooled in one location. Our results show that synergy conformal predictors based on training data randomly sampled with replacement can compete with other conformal setups, while using completely separate training sets often results in worse performance. However, in a federated setup where no method has access to all the data, synergy conformal prediction is shown to give promising results. Based on our study, we conclude that synergy conformal predictors are a valuable addition to the conformal prediction toolbox., (© 2021. The Author(s).)

Machine learning methods are widely used in drug discovery and toxicity prediction. While showing overall good performance in cross-validation studies, their predictive power (often) drops in cases where the query samples have drifted from the training data's descriptor space. Thus, the assumption for applying machine learning algorithms, that training and test data stem from the same distribution, might not always be fulfilled. In this work, conformal prediction is used to assess the calibration of the models. Deviations from the expected error may indicate that training and test data originate from different distributions. Exemplified on the Tox21 datasets, composed of chronologically released Tox21Train, Tox21Test and Tox21Score subsets, we observed that while internally valid models could be trained using cross-validation on Tox21Train, predictions on the external Tox21Score data resulted in higher error rates than expected. To improve the prediction on the external sets, a strategy exchanging the calibration set with more recent data, such as Tox21Test, has successfully been introduced. We conclude that conformal prediction can be used to diagnose data drifts and other issues related to model calibration. The proposed improvement strategy-exchanging the calibration data only-is convenient as it does not require retraining of the underlying model.

The reproducibility of experiments has been a long standing impediment for further scientific progress. Computational methods have been instrumental in drug discovery efforts owing to its multifaceted utilization for data collection, pre-processing, analysis and inference. This article provides an in-depth coverage on the reproducibility of computational drug discovery. This review explores the following topics: (1) the current state-of-the-art on reproducible research, (2) research documentation (e.g. electronic laboratory notebook, Jupyter notebook, etc.), (3) science of reproducible research (i.e. comparison and contrast with related concepts as replicability, reusability and reliability), (4) model development in computational drug discovery, (5) computational issues on model development and deployment, (6) use case scenarios for streamlining the computational drug discovery protocol. In computational disciplines, it has become common practice to share data and programming codes used for numerical calculations as to not only facilitate reproducibility, but also to foster collaborations (i.e. to drive the project further by introducing new ideas, growing the data, augmenting the code, etc.). It is therefore inevitable that the field of computational drug design would adopt an open approach towards the collection, curation and sharing of data/code.

Ligand-based predictive modeling is widely used to generate predictive models aiding decision making in e.g. drug discovery projects. With growing data sets and requirements on low modeling time comes the necessity to analyze data sets efficiently to support rapid and robust modeling. In this study we analyzed four data sets and studied the efficiency of machine learning methods on sparse data structures, utilizing Morgan fingerprints of different radii and hash sizes, and compared with molecular signatures descriptor of different height. We specifically evaluated the effect these parameters had on modeling time, predictive performance, and memory requirements using two implementations of random forest; Scikit-learn as well as FEST. We also compared with a support vector machine implementation. Our results showed that unhashed fingerprints yield significantly better accuracy than hashed fingerprints ([Formula: see text]), with no pronounced deterioration in modeling time and memory usage. Furthermore, the fast execution and low memory usage of the FEST algorithm suggest that it is a good alternative for large, high dimensional sparse data. Both support vector machines and random forest performed equally well but results indicate that the support vector machine was better at using the extra information from larger values of the Morgan fingerprint's radius.

Lipophilicity is a major determinant of ADMET properties and overall suitability of drug candidates. We have developed large-scale models to predict water-octanol distribution coefficient (logD) for chemical compounds, aiding drug discovery projects. Using ACD/logD data for 1.6 million compounds from the ChEMBL database, models are created and evaluated by a support-vector machine with a linear kernel using conformal prediction methodology, outputting prediction intervals at a specified confidence level. The resulting model shows a predictive ability of [Formula: see text] and with the best performing nonconformity measure having median prediction interval of [Formula: see text] log units at 80% confidence and [Formula: see text] log units at 90% confidence. The model is available as an online service via an OpenAPI interface, a web page with a molecular editor, and we also publish predictive values at 90% confidence level for 91 M PubChem structures in RDF format for download and as an URI resolver service.

Background: Docking and scoring large libraries of ligands against target proteins forms the basis of structure-based virtual screening. The problem is trivially parallelizable, and calculations are generally carried out on computer clusters or on large workstations in a brute force manner, by docking and scoring all available ligands., Contribution: In this study we propose a strategy that is based on iteratively docking a set of ligands to form a training set, training a ligand-based model on this set, and predicting the remainder of the ligands to exclude those predicted as 'low-scoring' ligands. Then, another set of ligands are docked, the model is retrained and the process is repeated until a certain model efficiency level is reached. Thereafter, the remaining ligands are docked or excluded based on this model. We use SVM and conformal prediction to deliver valid prediction intervals for ranking the predicted ligands, and Apache Spark to parallelize both the docking and the modeling., Results: We show on 4 different targets that conformal prediction based virtual screening (CPVS) is able to reduce the number of docked molecules by 62.61% while retaining an accuracy for the top 30 hits of 94% on average and a speedup of 3.7. The implementation is available as open source via GitHub ( https://github.com/laeeq80/spark-cpvs ) and can be run on high-performance computers as well as on cloud resources.

Background: The Chemistry Development Kit (CDK) is a widely used open source cheminformatics toolkit, providing data structures to represent chemical concepts along with methods to manipulate such structures and perform computations on them. The library implements a wide variety of cheminformatics algorithms ranging from chemical structure canonicalization to molecular descriptor calculations and pharmacophore perception. It is used in drug discovery, metabolomics, and toxicology. Over the last 10 years, the code base has grown significantly, however, resulting in many complex interdependencies among components and poor performance of many algorithms., Results: We report improvements to the CDK v2.0 since the v1.2 release series, specifically addressing the increased functional complexity and poor performance. We first summarize the addition of new functionality, such atom typing and molecular formula handling, and improvement to existing functionality that has led to significantly better performance for substructure searching, molecular fingerprints, and rendering of molecules. Second, we outline how the CDK has evolved with respect to quality control and the approaches we have adopted to ensure stability, including a code review mechanism., Conclusions: This paper highlights our continued efforts to provide a community driven, open source cheminformatics library, and shows that such collaborative projects can thrive over extended periods of time, resulting in a high-quality and performant library. By taking advantage of community support and contributions, we show that an open source cheminformatics project can act as a peer reviewed publishing platform for scientific computing software. Graphical abstract CDK 2.0 provides new features and improved performance.

Background: Structure-based virtual screening is an in-silico method to screen a target receptor against a virtual molecular library. Applying docking-based screening to large molecular libraries can be computationally expensive, however it constitutes a trivially parallelizable task. Most of the available parallel implementations are based on message passing interface, relying on low failure rate hardware and fast network connection. Google's MapReduce revolutionized large-scale analysis, enabling the processing of massive datasets on commodity hardware and cloud resources, providing transparent scalability and fault tolerance at the software level. Open source implementations of MapReduce include Apache Hadoop and the more recent Apache Spark., Results: We developed a method to run existing docking-based screening software on distributed cloud resources, utilizing the MapReduce approach. We benchmarked our method, which is implemented in Apache Spark, docking a publicly available target receptor against [Formula: see text]2.2 M compounds. The performance experiments show a good parallel efficiency (87%) when running in a public cloud environment., Conclusion: Our method enables parallel Structure-based virtual screening on public cloud resources or commodity computer clusters. The degree of scalability that we achieve allows for trying out our method on relatively small libraries first and then to scale to larger libraries. Our implementation is named Spark-VS and it is freely available as open source from GitHub (https://github.com/mcapuccini/spark-vs).Graphical abstract.

Predictive modelling in drug discovery is challenging to automate as it often contains multiple analysis steps and might involve cross-validation and parameter tuning that create complex dependencies between tasks. With large-scale data or when using computationally demanding modelling methods, e-infrastructures such as high-performance or cloud computing are required, adding to the existing challenges of fault-tolerant automation. Workflow management systems can aid in many of these challenges, but the currently available systems are lacking in the functionality needed to enable agile and flexible predictive modelling. We here present an approach inspired by elements of the flow-based programming paradigm, implemented as an extension of the Luigi system which we name SciLuigi. We also discuss the experiences from using the approach when modelling a large set of biochemical interactions using a shared computer cluster.Graphical abstract.

Xenobiotic metabolism is an active research topic but the limited amount of openly available high-quality biotransformation data constrains predictive modeling. Current database often default to commonly available information: which enzyme metabolizes a compound, but neither experimental conditions nor the atoms that undergo metabolization are captured. We present XMetDB, an open access database for drugs and other xenobiotics and their respective metabolites. The database contains chemical structures of xenobiotic biotransformations with substrate atoms annotated as reaction centra, the resulting product formed, and the catalyzing enzyme, type of experiment, and literature references. Associated with the database is a web interface for the submission and retrieval of experimental metabolite data for drugs and other xenobiotics in various formats, and a web API for programmatic access is also available. The database is open for data deposition, and a curation scheme is in place for quality control. An extensive guide on how to enter experimental data into is available from the XMetDB wiki. XMetDB formalizes how biotransformation data should be reported, and the openly available systematically labeled data is a big step forward towards better models for predictive metabolism.

The increasing size of datasets in drug discovery makes it challenging to build robust and accurate predictive models within a reasonable amount of time. In order to investigate the effect of dataset sizes on predictive performance and modelling time, ligand-based regression models were trained on open datasets of varying sizes of up to 1.2 million chemical structures. For modelling, two implementations of support vector machines (SVM) were used. Chemical structures were described by the signatures molecular descriptor. Results showed that for the larger datasets, the LIBLINEAR SVM implementation performed on par with the well-established libsvm with a radial basis function kernel, but with dramatically less time for model building even on modest computer resources. Using a non-linear kernel proved to be infeasible for large data sizes, even with substantial computational resources on a computer cluster. To deploy the resulting models, we extended the Bioclipse decision support framework to support models from LIBLINEAR and made our models of logD and solubility available from within Bioclipse.

Background: Making data available as Linked Data using Resource Description Framework (RDF) promotes integration with other web resources. RDF documents can natively link to related data, and others can link back using Uniform Resource Identifiers (URIs). RDF makes the data machine-readable and uses extensible vocabularies for additional information, making it easier to scale up inference and data analysis., Results: This paper describes recent developments in an ongoing project converting data from the ChEMBL database into RDF triples. Relative to earlier versions, this updated version of ChEMBL-RDF uses recently introduced ontologies, including CHEMINF and CiTO; exposes more information from the database; and is now available as dereferencable, linked data. To demonstrate these new features, we present novel use cases showing further integration with other web resources, including Bio2RDF, Chem2Bio2RDF, and ChemSpider, and showing the use of standard ontologies for querying., Conclusions: We have illustrated the advantages of using open standards and ontologies to link the ChEMBL database to other databases. Using those links and the knowledge encoded in standards and ontologies, the ChEMBL-RDF resource creates a foundation for integrated semantic web cheminformatics applications, such as the presented decision support.

Background: The InChI algorithms are written in C++ and not available as Java library. Integration into software written in Java therefore requires a bridge between C and Java libraries, provided by the Java Native Interface (JNI) technology., Results: We here describe how the InChI library is used in the Bioclipse workbench and the Chemistry Development Kit (CDK) cheminformatics library. To make this possible, a JNI bridge to the InChI library was developed, JNI-InChI, allowing Java software to access the InChI algorithms. By using this bridge, the CDK project packages the InChI binaries in a module and offers easy access from Java using the CDK API. The Bioclipse project packages and offers InChI as a dynamic OSGi bundle that can easily be used by any OSGi-compliant software, in addition to the regular Java Archive and Maven bundles. Bioclipse itself uses the InChI as a key component and calculates it on the fly when visualizing and editing chemical structures. We demonstrate the utility of InChI with various applications in CDK and Bioclipse, such as decision support for chemical liability assessment, tautomer generation, and for knowledge aggregation using a linked data approach., Conclusions: These results show that the InChI library can be used in a variety of Java library dependency solutions, making the functionality easily accessible by Java software, such as in the CDK. The applications show various ways the InChI has been used in Bioclipse, to enrich its functionality.