An automated framework for QSAR model building

Kausar and Falcao J Cheminform (2018) 10:1 https://doi.org/10.1186/s13321-017-0256-5 RESEARCH ARTICLE Open Access An automated framework for QSAR model building Samina Kausar1,2 and Andre O. Falcao1,2* Abstract Background: In-silico quantitative structure–activity relationship (QSAR) models based tools are widely used to screen huge databases of compounds in order to determine the biological properties of chemical molecules based on their chemical structure. With the passage of time, the exponentially growing amount of synthesized and known chemicals data demands computationally efficient automated QSAR modeling tools, available to researchers that may lack extensive knowledge of machine learning modeling. Thus, a fully automated and advanced modeling platform can be an important addition to the QSAR community. Results: In the presented workflow the process from data preparation to model building and validation has been completely automated. The most critical modeling tasks (data curation, data set characteristics evaluation, variable selection and validation) that largely influence the performance of QSAR models were focused. It is also included the ability to quickly evaluate the feasibility of a given data set to be modeled. The developed framework is tested on data sets of thirty different problems. The best-optimized feature selection methodology in the developed workflow is able to remove 62–99% of all redundant data. On average, about 19% of the prediction error was reduced by using feature selection producing an increase of 49% in the percentage of variance explained (PVE) compared to models without feature selection. Selecting only the models with a modelability score above 0.6, average PVE scores were 0.71. A strong correlation was verified between the modelability scores and the PVE of the models produced with variable selection. Conclusions: We developed an extendable and highly customizable fully automated QSAR modeling framework. This designed workflow does not require any advanced parameterization nor depends on users decisions or expertise in machine learning/programming. With just a given target or problem, the workflow follows an unbiased standard protocol to develop reliable QSAR models by directly accessing online manually curated databases or by using private data sets. The other distinctive features of the workflow include prior estimation of data modelability to avoid time-consuming modeling trials for non modelable data sets, an efficient variable selection procedure and the facility of output availability at each modeling task for the diverse application and reproduction of historical predictions. The results reached on a selection of thirty QSAR problems suggest that the approach is capable of building reliable models even for challenging problems. Keywords: Quantitative structure–activity relationship (QSAR), Machine learning, Feature selection, Variable importance, Random forests, Support vector machines, KNIME, Data set modelability *Correspondence: 1 LaSIGE, Departamento de Informática, Faculdade de Ciências, Universidade de Lisboa, 1749‑016 Lisbon, Portugal Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Kausar and Falcao J Cheminform (2018) 10:1 Introduction Background The advantages of automation of repetitive tasks in the laborious drug discovery process are numerous and include increased research quality by reducing error along with significant time saving, boosted up productivity, and capacity to name a few. In this era where large amounts of data are produced every day and large computational resources are available, the introduction of machine learning approaches has significantly automated the drug discovery procedure and provides a faster alternative for ultrahigh-throughput screening of large databases of chemical molecules against a biological target [1–3]. Machine learning approaches are being applied in the drug discovery cycle to produce a robust model, capable of empirical predictions of biological properties of candidate compounds for new therapeutic molecules. Many successful studies have been reported in the literature which attests the importance of machine learning approaches combined with traditional practices to approach medicinal chemistry challenges [4]. In traditional lab work methodologies, many expensive tests are often required which many times include animal testing to provide information about human safety for suggested chemicals. The legislation does not support such frequent experiments on laboratory animals, but rather promotes the sharing of data to the use of integrated alternative in-vitro and in-silico strategies of toxicokinetics [5–7]. Currently the Avicenna Research and Technological Roadmap, funded by the European Commission, strongly suggests the use of in-silico techniques coupled with clinical trials [8]. This framework describes strategic priorities to establish the safety assessment of new medical interventions and at the same time minimizes the ethically concerned activities such as the animal or human experimentation. Several available in-silico QSAR models based tools are widely used to screen very large databases of compounds in order to determine toxicity or any desired biological effects of chemical molecules based on their chemical structure [9, 10]. The well-characterized internationally accepted validation principles for creating validated models have been used by regulatory agencies of United Sates (US) and gaining a boost in the European Union (EU) too [8, 11–13]. In the EU, the standard recommendations of chemicals risk assessment by regulatory QSAR models has been set by the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) [14] and the Organization for Economic Cooperation and Development (OECD) [15]. The progress of such projects highlights the increased importance of Page 2 of 23 productivity gains from fully accessible automation in the drug discovery and QSAR modeling fields. These days, the aim of pharmaceutical projects is the integration of complex non-homogeneous data to build global models intended to be applicable within wide ranges of chemical space. However, with the passage of time, there is an exponentially growing amount of synthesized and known chemical compounds data being added to the many existing molecule data (...truncated)