GA(M)E-QSAR: A Novel, Fully Automatic Genetic-Algorithm-(Meta)-Ensembles Approach for Binary Classification in Ligand-Based Drug Design

Prediction of blood-brain barrier permeability using machine learning approaches based on various molecular representation

The assessment of compound blood-brain barrier (BBB) permeability poses a significant challenge in the discovery of drugs targeting the central nervous system. Conventional experimental approaches to measure BBB permeability are labor-intensive, cost-ineffective, and time-consuming. In this study, we constructed six machine learning classification models by combining various machine learning algorithms and molecular representations. The model based on ExtraTree algorithm and random partitioning strategy obtains the best prediction result, with AUC value of 0.932±0.004 and balanced accuracy (BA) of 0.837±0.010 for the test set. We employed the SHAP method to identify important features associated with BBB permeability. In addition, matched molecular pair (MMP) analysis and representative substructure derivation method were utilized to uncover the transformation rules and distinctive structural features of BBB permeable compounds. The machine learning models proposed in this work can serve as an effective tool for assessing BBB permeability in the drug discovery for central nervous system disease.

Prediction of compound-target interaction using several artificial intelligence algorithms and comparison with a consensus-based strategy

For understanding a chemical compound's mechanism of action and its side effects, as well as for drug discovery, it is crucial to predict its possible protein targets. This study examines 15 developed target-centric models (TCM) employing different molecular descriptions and machine learning algorithms. They were contrasted with 17 third-party models implemented as web tools (WTCM). In both sets of models, consensus strategies were implemented as potential improvement over individual predictions. The findings indicate that TCM reach f1-score values greater than 0.8. Comparing both approaches, the best TCM achieves values of 0.75, 0.61, 0.25 and 0.38 for true positive/negative rates (TPR, TNR) and false negative/positive rates (FNR, FPR); outperforming the best WTCM. Moreover, the consensus strategy proves to have the most relevant results in the top 20 % of target profiles. TCM consensus reach TPR and FNR values of 0.98 and 0; while on WTCM reach values of 0.75 and 0.24. The implemented computational tool with the TCM and their consensus strategy at: https://bioquimio.udla.edu.ec/tidentification01/ . Scientific Contribution: We compare and discuss the performances of 17 public compound-target interaction prediction models and 15 new constructions. We also explore a compound-target interaction prioritization strategy using a consensus approach, and we analyzed the challenging involved in interactions modeling.

Improvement of Prediction Performance With Conjoint Molecular Fingerprint in Deep Learning

The accurate predicting of physical properties and bioactivity of drug molecules in deep learning depends on how molecules are represented. Many types of molecular descriptors have been developed for quantitative structure-activity/property relationships quantitative structure-activity relationships (QSPR). However, each molecular descriptor is optimized for a specific application with encoding preference. Considering that standalone featurization methods may only cover parts of information of the chemical molecules, we proposed to build the conjoint fingerprint by combining two supplementary fingerprints. The impact of conjoint fingerprint and each standalone fingerprint on predicting performance was systematically evaluated in predicting the logarithm of the partition coefficient (logP) and binding affinity of protein-ligand by using machine learning/deep learning (ML/DL) methods, including random forest (RF), support vector regression (SVR), extreme gradient boosting (XGBoost), long short-term memory network (LSTM), and deep neural network (DNN). The results demonstrated that the conjoint fingerprint yielded improved predictive performance, even outperforming the consensus model using two standalone fingerprints among four out of five examined methods. Given that the conjoint fingerprint scheme shows easy extensibility and high applicability, we expect that the proposed conjoint scheme would create new opportunities for continuously improving predictive performance of deep learning by harnessing the complementarity of various types of fingerprints.

Ensemble-Based Modeling of Chemical Compounds with Antimalarial Activity

Background:

Malaria or Paludism is a tropical disease caused by parasites of the Plasmodium genre and transmitted to humans through the bite of infected mosquitos of the Anopheles genre. This pathology is considered one of the first causes of death in tropical countries and, despite several existing therapies, they have a high toxicity. Computational methods based on Quantitative Structure- Activity Relationship studies have been widely used in drug design work flows.

Objective:

The main goal of the current research is to develop computational models for the identification of antimalarial hit compounds.

Materials and Methods:

For this, a data set suitable for the modeling of the antimalarial activity of chemical compounds was compiled from the literature and subjected to a thorough curation process. In addition, the performance of a diverse set of ensemble-based classification methodologies was evaluated and one of these ensembles was selected as the most suitable for the identification of antimalarial hits based on its virtual screening performance. Data curation was conducted to minimize noise. Among the explored ensemble-based methods, the one combining Genetic Algorithms for the selection of the base classifiers and Majority Vote for their aggregation showed the best performance.

Results:

Our results also show that ensemble modeling is an effective strategy for the QSAR modeling of highly heterogeneous datasets in the discovery of potential antimalarial compounds.

Conclusion:

It was determined that the best performing ensembles were those that use Genetic Algorithms as a method of selection of base models and Majority Vote as the aggregation method.

A combined drug discovery strategy based on machine learning and molecular docking

Data mining methods based on machine learning play an increasingly important role in drug design and discovery. In the current work, eight machine learning methods including decision trees, k-Nearest neighbor, support vector machines, random forests, extremely randomized trees, AdaBoost, gradient boosting trees, and XGBoost were evaluated comprehensively through a case study of ACC inhibitor data sets. Internal and external data sets were employed for cross-validation of the eight machine learning methods. Results showed that the extremely randomized trees model performed best and was adopted as the first step of virtual screening. Together with structure-based virtual screening in the second step, this combined strategy obtained desirable results. This work indicates that the combination of machine learning methods with traditional structure-based virtual screening can effectively strengthen the ability in finding potential hits from large compound database for a given target.

A desirability-based multi objective approach for the virtual screening discovery of broad-spectrum anti-gastric cancer agents

Gastric cancer is the third leading cause of cancer-related mortality worldwide and despite advances in prevention, diagnosis and therapy, it is still regarded as a global health concern. The efficacy of the therapies for gastric cancer is limited by a poor response to currently available therapeutic regimens. One of the reasons that may explain these poor clinical outcomes is the highly heterogeneous nature of this disease. In this sense, it is essential to discover new molecular agents capable of targeting various gastric cancer subtypes simultaneously. Here, we present a multi-objective approach for the ligand-based virtual screening discovery of chemical compounds simultaneously active against the gastric cancer cell lines AGS, NCI-N87 and SNU-1. The proposed approach relays in a novel methodology based on the development of ensemble models for the bioactivity prediction against each individual gastric cancer cell line. The methodology includes the aggregation of one ensemble per cell line using a desirability-based algorithm into virtual screening protocols. Our research leads to the proposal of a multi-targeted virtual screening protocol able to achieve high enrichment of known chemicals with anti-gastric cancer activity. Specifically, our results indicate that, using the proposed protocol, it is possible to retrieve almost 20 more times multi-targeted compounds in the first 1% of the ranked list than what is expected from a uniform distribution of the active ones in the virtual screening database. More importantly, the proposed protocol attains an outstanding initial enrichment of known multi-targeted anti-gastric cancer agents.

Systemic QSAR and phenotypic virtual screening: chasing butterflies in drug discovery

Current advances in systems biology suggest a new change of paradigm reinforcing the holistic nature of the drug discovery process. According to the principles of systems biology, a simple drug perturbing a network of targets can trigger complex reactions. Therefore, it is possible to connect initial events with final outcomes and consequently prioritize those events, leading to a desired effect. Here, we introduce a new concept, 'Systemic Chemogenomics/Quantitative Structure-Activity Relationship (QSAR)'. To elaborate on the concept, relevant information surrounding it is addressed. The concept is challenged by implementing a systemic QSAR approach for phenotypic virtual screening (VS) of candidate ligands acting as neuroprotective agents in Parkinson's disease (PD). The results support the suitability of the approach for the phenotypic prioritization of drug candidates.

An Introduction to the Basic Concepts in QSAR-Aided Drug Design

The need for the development of new drugs to combat existing and newly identified conditions is unavoidable. One of the important tools used in the advanced drug development pipeline is computer-aided drug design. Traditionally, to find a drug many ligands were synthesized and evaluated for their effectiveness using suitable bioassays and if all other drug-likeness features were met, the candidate(s) would possibly reach the market. Although this approach is still in use in advanced format, computational methods are an indispensable component of modern drug development projects. One of the methods used from very early days of rationalizing the drug design approaches is Quantitative Structure-Activity Relationship (QSAR). This chapter overviews QSAR modeling steps by introducing molecular descriptors, mathematical model development for relating biological activities to molecular structures, and model validation. At the end, several successful cases where QSAR studies were used extensively are presented.

Profiling the interaction mechanism of indole-based derivatives targeting the HIV-1 gp120 receptor

A glycoprotein exposed on a viral surface, human immunodeficiency virus type 1 (HIV-1) gp120 is essential for virus entry into cells as it plays a vital role in seeking out specific cell surface receptors for entry.

An Introduction to the Basic Concepts in QSAR-Aided Drug Design

The need for the development of new drugs to combat existing and newly identified conditions is unavoidable. One of the important tools used in the advanced drug development pipeline is computer-aided drug design. Traditionally, to find a drug many ligands were synthesized and evaluated for their effectiveness using suitable bioassays and if all other drug-likeness features were met, the candidate(s) would possibly reach the market. Although this approach is still in use in advanced format, computational methods are an indispensable component of modern drug development projects. One of the methods used from very early days of rationalizing the drug design approaches is Quantitative Structure-Activity Relationship (QSAR). This chapter overviews QSAR modeling steps by introducing molecular descriptors, mathematical model development for relating biological activities to molecular structures, and model validation. At the end, several successful cases where QSAR studies were used extensively are presented.

AutoWeka: toward an automated data mining software for QSAR and QSPR studies

In biology and chemistry, a key goal is to discover novel compounds affording potent biological activity or chemical properties. This could be achieved through a chemical intuition-driven trial-and-error process or via data-driven predictive modeling. The latter is based on the concept of quantitative structure-activity/property relationship (QSAR/QSPR) when applied in modeling the biological activity and chemical properties, respectively, of compounds. Data mining is a powerful technology underlying QSAR/QSPR as it harnesses knowledge from large volumes of high-dimensional data via multivariate analysis. Although extremely useful, the technicalities of data mining may overwhelm potential users, especially those in the life sciences. Herein, we aim to lower the barriers to access and utilization of data mining software for QSAR/QSPR studies. AutoWeka is an automated data mining software tool that is powered by the widely used machine learning package Weka. The software provides a user-friendly graphical interface along with an automated parameter search capability. It employs two robust and popular machine learning methods: artificial neural networks and support vector machines. This chapter describes the practical usage of AutoWeka and relevant tools in the development of predictive QSAR/QSPR models.

AVAILABILITY

The software is freely available at http://www.mt.mahidol.ac.th/autoweka.

Toward the computer-aided discovery of FabH inhibitors. Do predictive QSAR models ensure high quality virtual screening performance?

Antibiotic resistance has increased over the past two decades. New approaches for the discovery of novel antibacterials are required and innovative strategies will be necessary to identify novel and effective candidates. Related to this problem, the exploration of bacterial targets that remain unexploited by the current antibiotics in clinical use is required. One of such targets is the β-ketoacyl-acyl carrier protein synthase III (FabH). Here, we report a ligand-based modeling methodology for the virtual-screening of large collections of chemical compounds in the search of potential FabH inhibitors. QSAR models are developed for a diverse dataset of 296 FabH inhibitors using an in-house modeling framework. All models showed high fitting, robustness, and generalization capabilities. We further investigated the performance of the developed models in a virtual screening scenario. To carry out this investigation, we implemented a desirability-based algorithm for decoys selection that was shown effective in the selection of high quality decoys sets. Once the QSAR models were validated in the context of a virtual screening experiment their limitations arise. For this reason, we explored the potential of ensemble modeling to overcome the limitations associated to the use of single classifiers. Through a detailed evaluation of the virtual screening performance of ensemble models it was evidenced, for the first time to our knowledge, the benefits of this approach in a virtual screening scenario. From all the obtained results, we could arrive to a significant main conclusion: at least for FabH inhibitors, virtual screening performance is not guaranteed by predictive QSAR models.

Classifier ensemble based on feature selection and diversity measures for predicting the affinity of A(2B) adenosine receptor antagonists

A(2B) adenosine receptor antagonists may be beneficial in treating diseases like asthma, diabetes, diabetic retinopathy, and certain cancers. This has stimulated research for the development of potent ligands for this subtype, based on quantitative structure-affinity relationships. In this work, a new ensemble machine learning algorithm is proposed for classification and prediction of the ligand-binding affinity of A(2B) adenosine receptor antagonists. This algorithm is based on the training of different classifier models with multiple training sets (composed of the same compounds but represented by diverse features). The k-nearest neighbor, decision trees, neural networks, and support vector machines were used as single classifiers. To select the base classifiers for combining into the ensemble, several diversity measures were employed. The final multiclassifier prediction results were computed from the output obtained by using a combination of selected base classifiers output, by utilizing different mathematical functions including the following: majority vote, maximum and average probability. In this work, 10-fold cross- and external validation were used. The strategy led to the following results: i) the single classifiers, together with previous features selections, resulted in good overall accuracy, ii) a comparison between single classifiers, and their combinations in the multiclassifier model, showed that using our ensemble gave a better performance than the single classifier model, and iii) our multiclassifier model performed better than the most widely used multiclassifier models in the literature. The results and statistical analysis demonstrated the supremacy of our multiclassifier approach for predicting the affinity of A(2B) adenosine receptor antagonists, and it can be used to develop other QSAR models.