CompScore: Boosting Structure-Based Virtual Screening Performance by Incorporating Docking Scoring Function Components into Consensus Scoring.

Consensus scoring has become a commonly used strategy within structure-based virtual screening (VS) workflows with improved performance compared to those based in a single scoring function. However, no research has been devoted to analyze the worth of docking scoring functions components in consensus scoring. We implemented and tested a method that incorporates docking scoring functions components into the setting of high performance VS workflows. This method uses genetic algorithms for finding the combination of scoring components that maximizes the VS enrichment for any target. Our methodology was validated using a data set including ligands and decoys for 102 targets that have been widely used in VS validation studies. Results show that our approach outperforms other methods for all targets. It also boosts the initial enrichment performance of the traditional use of whole scoring functions in consensus scoring by an average of 45%. Our methodology showed to be outstandingly predictive when challenged to rescore external (previously unseen) data. Remarkably, CompScore was able not only to retain its performance after redocking with a different software, but also proved that the enrichment obtained was not artificial. CompScore is freely available at: http://bioquimio.udla.edu.ec/compscore/ .

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.

Net-Net Auto Machine Learning (AutoML) Prediction of Complex Ecosystems.

Biological Ecosystem Networks (BENs) are webs of biological species (nodes) establishing trophic relationships (links). Experimental confirmation of all possible links is difficult and generates a huge volume of information. Consequently, computational prediction becomes an important goal. Artificial Neural Networks (ANNs) are Machine Learning (ML) algorithms that may be used to predict BENs, using as input Shannon entropy information measures (Shk) of known ecosystems to train them. However, it is difficult to select a priori which ANN topology will have a higher accuracy. Interestingly, Auto Machine Learning (AutoML) methods focus on the automatic selection of the more efficient ML algorithms for specific problems. In this work, a preliminary study of a new approach to AutoML selection of ANNs is proposed for the prediction of BENs. We call it the Net-Net AutoML approach, because it uses for the first time Shk values of both networks involving BENs (networks to be predicted) and ANN topologies (networks to be tested). Twelve types of classifiers have been tested for the Net-Net model including linear, Bayesian, trees-based methods, multilayer perceptrons and deep neuronal networks. The best Net-Net AutoML model for 338,050 outputs of 10 ANN topologies for links of 69 BENs was obtained with a deep fully connected neuronal network, characterized by a test accuracy of 0.866 and a test AUROC of 0.935. This work paves the way for the application of Net-Net AutoML to other systems or ML algorithms.

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.

Consensus strategy in genes prioritization and combined bioinformatics analysis for preeclampsia pathogenesis.

BACKGROUND: Preeclampsia is a multifactorial disease with unknown pathogenesis. Even when recent studies explored this disease using several bioinformatics tools, the main objective was not directed to pathogenesis. Additionally, consensus prioritization was proved to be highly efficient in the recognition of genes-disease association. However, not information is available about the consensus ability to early recognize genes directly involved in pathogenesis. Therefore our aim in this study is to apply several theoretical approaches to explore preeclampsia; specifically those genes directly involved in the pathogenesis. METHODS: We firstly evaluated the consensus between 12 prioritization strategies to early recognize pathogenic genes related to preeclampsia. A communality analysis in the protein-protein interaction network of previously selected genes was done including further enrichment analysis. The enrichment analysis includes metabolic pathways as well as gene ontology. Microarray data was also collected and used in order to confirm our results or as a strategy to weight the previously enriched pathways. RESULTS: The consensus prioritized gene list was rationally filtered to 476 genes using several criteria. The communality analysis showed an enrichment of communities connected with VEGF-signaling pathway. This pathway is also enriched considering the microarray data. Our result point to VEGF, FLT1 and KDR as relevant pathogenic genes, as well as those connected with NO metabolism. CONCLUSION: Our results revealed that consensus strategy improve the detection and initial enrichment of pathogenic genes, at least in preeclampsia condition. Moreover the combination of the first percent of the prioritized genes with protein-protein interaction network followed by communality analysis reduces the gene space. This approach actually identifies well known genes related with pathogenesis. However, genes like HSP90, PAK2, CD247 and others included in the first 1% of the prioritized list need to be further explored in preeclampsia pathogenesis through experimental approaches.

From flamingo dance to (desirable) drug discovery: a nature-inspired approach.

The therapeutic effects of drugs are well known to result from their interaction with multiple intracellular targets. Accordingly, the pharma industry is currently moving from a reductionist approach based on a 'one-target fixation' to a holistic multitarget approach. However, many drug discovery practices are still procedural abstractions resulting from the attempt to understand and address the action of biologically active compounds while preventing adverse effects. Here, we discuss how drug discovery can benefit from the principles of evolutionary biology and report two real-life case studies. We do so by focusing on the desirability principle, and its many features and applications, such as machine learning-based multicriteria virtual screening.

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.

Chemoinformatics Profiling of the Chromone Nucleus as a MAO-B/A2AAR Dual Binding Scaffold.

BACKGROUND: In the context of the current drug discovery efforts to find disease modifying therapies for Parkinson's disease (PD) the current single target strategy has proved inefficient. Consequently, the search for multi-potent agents is attracting more and more attention due to the multiple pathogenetic factors implicated in PD. Multiple evidences points to the dual inhibition of the monoamine oxidase B (MAO-B), as well as adenosine A2A receptor (A2AAR) blockade, as a promising approach to prevent the neurodegeneration involved in PD. Currently, only two chemical scaffolds has been proposed as potential dual MAO-B inhibitors/A2AAR antagonists (caffeine derivatives and benzothiazinones). METHODS: In this study, we conduct a series of chemoinformatics analysis in order to evaluate and advance the potential of the chromone nucleus as a MAO-B/A2AAR dual binding scaffold. RESULTS: The information provided by SAR data mining analysis based on network similarity graphs and molecular docking studies support the suitability of the chromone nucleus as a potential MAOB/ A2AAR dual binding scaffold. Additionally, a virtual screening tool based on a group fusion similarity search approach was developed for the prioritization of potential MAO-B/A2AAR dual binder candidates. Among several data fusion schemes evaluated, the MEAN-SIM and MIN-RANK GFSS approaches demonstrated to be efficient virtual screening tools. Then, a combinatorial library potentially enriched with MAO-B/A2AAR dual binding chromone derivatives was assembled and sorted by using the MIN-RANK and then the MEAN-SIM GFSS VS approaches. CONCLUSION: The information and tools provided in this work represent valuable decision making elements in the search of novel chromone derivatives with a favorable dual binding profile as MAOB inhibitors and A2AAR antagonists with the potential to act as a disease-modifying therapeutic for Parkinson's disease.

Fusing Docking Scoring Functions Improves the Virtual Screening Performance for Discovering Parkinson's Disease Dual Target Ligands.

BACKGROUND: Virtual methodologies have become essential components of the drug discovery pipeline. Specifically, structure-based drug design methodologies exploit the 3D structure of molecular targets to discover new drug candidates through molecular docking. Recently, dual target ligands of the Adenosine A2A Receptor and Monoamine Oxidase B enzyme have been proposed as effective therapies for the treatment of Parkinson's disease. METHODS: In this paper we propose a structure-based methodology, which is extensively validated, for the discovery of dual Adenosine A2A Receptor/Monoamine Oxidase B ligands. This methodology involves molecular docking studies against both receptors and the evaluation of different scoring functions fusion strategies for maximizing the initial virtual screening enrichment of known dual ligands. RESULTS: The developed methodology provides high values of enrichment of known ligands, which outperform that of the individual scoring functions. At the same time, the obtained ensemble can be translated in a sequence of steps that should be followed to maximize the enrichment of dual target Adenosine A2A Receptor antagonists and Monoamine Oxidase B inhibitors. CONCLUSION: Information relative to docking scores to both targets have to be combined for achieving high dual ligands enrichment. Combining the rankings derived from different scoring functions proved to be a valuable strategy for improving the enrichment relative to single scoring function in virtual screening experiments.

Generalized Molecular Descriptors Derived From Event-Based Discrete Derivative.

In the present study, a generalized approach for molecular structure characterization is introduced, based on the relation frequency matrix (F) representation of the molecular graph and the subsequent calculation of the corresponding discrete derivative (finite difference) over a pair of elements (atoms). In earlier publications (22- 24), an unique event, named connected subgraphs, (based on the Kier-Hall's subgraphs) was systematically employed for the computation of the matrix F. The present report is a generalization of this notion, in which eleven additional events are introduced, classified in three categories, namely, topological (terminal paths, vertex path incidence, quantum subgraphs, walks of length k, Sach's subgraphs), fingerprints (MACCs, E-state and substructure fingerprints) and atomic contributions (Ghose and Crippen atom-types for hydrophobicity and refractivity) for F generation. The events are intended to capture diverse information by the generation or search of different kinds of substructures from the graph representation of a molecule. The discrete derivative over duplex atom relations are calculated for each event, and the resulting derivatives, local vertex invariants (LOVIs) are finally obtained. These LOVIs are subsequently employed as the basis for the calculation of global and local indices over groups of atoms (heteroatoms, halogens, methyl carbons, etc.), by using norms, means, statistics and classical algorithms as aggregator (fusion) operators. These indices were implemented in our house software DIVATI (Derivative Type Indices, a new module of TOMOCOMDCARDD system). DIVATI provides a friendly and cross-platform graphical user interface, developed in the Java programming language and is freely available at: http: //www.tomocomd.com. Factor analysis shows that the presented events are rather orthogonal and collect diverse information about the chemical structure. Finally, QSPR models were built to describe the logP and logK of 34 furylethylenes derivatives using the eleven events. Generally, the equations obtained according to these events showed high correlations, with the Sach's sub-graphs and Multiplicity events showing the best behavior in the description of logK (Q2 LOO value of 99.06%) and logP (Q2 LOO value of 98.1 %), respectively. These results show that these new eventbased indices constitute a powerful approach for chemoinformatics studies.

Physico-Chemical and Structural Interpretation of Discrete Derivative Indices on N-Tuples Atoms.

This report examines the interpretation of the Graph Derivative Indices (GDIs) from three different perspectives (i.e., in structural, steric and electronic terms). It is found that the individual vertex frequencies may be expressed in terms of the geometrical and electronic reactivity of the atoms and bonds, respectively. On the other hand, it is demonstrated that the GDIs are sensitive to progressive structural modifications in terms of: size, ramifications, electronic richness, conjugation effects and molecular symmetry. Moreover, it is observed that the GDIs quantify the interaction capacity among molecules and codify information on the activation entropy. A structure property relationship study reveals that there exists a direct correspondence between the individual frequencies of atoms and Hückel's Free Valence, as well as between the atomic GDIs and the chemical shift in NMR, which collectively validates the theory that these indices codify steric and electronic information of the atoms in a molecule. Taking in consideration the regularity and coherence found in experiments performed with the GDIs, it is possible to say that GDIs possess plausible interpretation in structural and physicochemical terms.

Probing the Hypothesis of SAR Continuity Restoration by the Removal of Activity Cliffs Generators in QSAR.

In this work we report the first attempt to study the effect of activity cliffs over the generalization ability of machine learning (ML) based QSAR classifiers, using as study case a previously reported diverse and noisy dataset focused on drug induced liver injury (DILI) and more than 40 ML classification algorithms. Here, the hypothesis of structure-activity relationship (SAR) continuity restoration by activity cliffs removal is tested as a potential solution to overcome such limitation. Previously, a parallelism was established between activity cliffs generators (ACGs) and instances that should be misclassified (ISMs), a related concept from the field of machine learning. Based on this concept we comparatively studied the classification performance of multiple machine learning classifiers as well as the consensus classifier derived from predictive classifiers obtained from training sets including or excluding ACGs. The influence of the removal of ACGs from the training set over the virtual screening performance was also studied for the respective consensus classifiers algorithms. In general terms, the removal of the ACGs from the training process slightly decreased the overall accuracy of the ML classifiers and multi-classifiers, improving their sensitivity (the weakest feature of ML classifiers trained with ACGs) but decreasing their specificity. Although these results do not support a positive effect of the removal of ACGs over the classification performance of ML classifiers, the "balancing effect" of ACG removal demonstrated to positively influence the virtual screening performance of multi-classifiers based on valid base ML classifiers. Specially, the early recognition ability was significantly favored after ACGs removal. The results presented and discussed in this work represent the first step towards the application of a remedial solution to the activity cliffs problem in QSAR studies.

Ligand-Based Virtual Screening Using Tailored Ensembles: A Prioritization Tool for Dual A2AAdenosine Receptor Antagonists / Monoamine Oxidase B Inhibitors.

BACKGROUND: Virtual Screening methodologies have emerged as efficient alternatives for the discovery of new drug candidates. At the same time, ensemble methods are nowadays frequently used to overcome the limitations of employing a single model in ligand-based drug design. However, many applications of ensemble methods to this area do not consider important aspects related to both virtual screening and the modeling process. During the application of ensemble methods to virtual screening the proper validation of the models in virtual screening conditions is often neglected. No analysis of the diversity of the ensemble members is performed frequently or no considerations regarding the applicability domain of the base models are being made. METHODS: In this research, we review basic concepts and definitions related to virtual screening. We comment recent applications of ensemble methods to ligand-based virtual screening and highlight their advantages and limitations. RESULTS: Next, we propose a method based on genetic algorithms optimization for the generation of virtual screening tailored ensembles which address the previously identified problems in the current applications of ensemble methods to virtual screening. CONCLUSION: Finally, the proposed methodology is successfully applied to the generation of ensemble models for the ligand-based virtual screening of dual target A2A adenosine receptor antagonists and MAO-B inhibitors as potential Parkinson's disease therapeutics.

Efficient and biologically relevant consensus strategy for Parkinson's disease gene prioritization.

BACKGROUND: The systemic information enclosed in microarray data encodes relevant clues to overcome the poorly understood combination of genetic and environmental factors in Parkinson's disease (PD), which represents the major obstacle to understand its pathogenesis and to develop disease-modifying therapeutics. While several gene prioritization approaches have been proposed, none dominate over the rest. Instead, hybrid approaches seem to outperform individual approaches. METHODS: A consensus strategy is proposed for PD related gene prioritization from mRNA microarray data based on the combination of three independent prioritization approaches: Limma, machine learning, and weighted gene co-expression networks. RESULTS: The consensus strategy outperformed the individual approaches in terms of statistical significance, overall enrichment and early recognition ability. In addition to a significant biological relevance, the set of 50 genes prioritized exhibited an excellent early recognition ability (6 of the top 10 genes are directly associated with PD). 40 % of the prioritized genes were previously associated with PD including well-known PD related genes such as SLC18A2, TH or DRD2. Eight genes (CCNH, DLK1, PCDH8, SLIT1, DLD, PBX1, INSM1, and BMI1) were found to be significantly associated to biological process affected in PD, representing potentially novel PD biomarkers or therapeutic targets. Additionally, several metrics of standard use in chemoinformatics are proposed to evaluate the early recognition ability of gene prioritization tools. CONCLUSIONS: The proposed consensus strategy represents an efficient and biologically relevant approach for gene prioritization tasks providing a valuable decision-making tool for the study of PD pathogenesis and the development of disease-modifying PD therapeutics.

Harmonization of QSAR Best Practices and Molecular Docking Provides an Efficient Virtual Screening Tool for Discovering New G-Quadruplex Ligands.

Telomeres and telomerase are key players in tumorogenesis. Among the various strategies proposed for telomerase inhibition or telomere uncapping, the stabilization of telomeric G-quadruplex (G4) structures is a very promising one. Additionally, G4 stabilizing ligands also act over tumors mediated by the alternative elongation of telomeres. Accordingly, the discovery of novel compounds able to act on telomeres and/or inhibit the telomerase enzyme by stabilizing DNA telomeric G4 structures as well as the development of approaches efficiently prioritizing such compounds constitute active areas of research in computational medicinal chemistry and anticancer drug discovery. In this direction, we applied a virtual screening strategy based on the rigorous application of QSAR best practices and its harmonized integration with structure-based methods. More than 600,000 compounds from commercial databases were screened, the first 99 compounds were prioritized, and 21 commercially available and structurally diverse candidates were purchased and submitted to experimental assays. Such strategy proved to be highly efficient in the prioritization of G4 stabilizer hits, with a hit rate of 23.5%. The best G4 stabilizer hit found exhibited a shift in melting temperature from FRET assay of +7.3 °C at 5 µM, while three other candidates also exhibited a promising stabilizing profile. The two most promising candidates also exhibited a good telomerase inhibitory ability and a mild inhibition of HeLa cells growth. None of these candidates showed antiproliferative effects in normal fibroblasts. Finally, the proposed virtual screening strategy proved to be a practical and reliable tool for the discovery of novel G4 ligands which can be used as starting points of further optimization campaigns.

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 ß-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.

Activity cliffs in drug discovery: Dr Jekyll or Mr Hyde?

The impact activity cliffs have on drug discovery is double-edged. For instance, whereas medicinal chemists can take advantage of regions in chemical space rich in activity cliffs, QSAR practitioners need to escape from such regions. The influence of activity cliffs in medicinal chemistry applications is extensively documented. However, the 'dark side' of activity cliffs (i.e. their detrimental effect on the development of predictive machine learning algorithms) has been understudied. Similarly, limited amounts of work have been devoted to propose potential solutions to the drawbacks of activity cliffs in similarity-based approaches. In this review, the duality of activity cliffs in medicinal chemistry and computational approaches is addressed, with emphasis on the rationale and potential solutions for handling the 'ugly face' of activity cliffs.

Chemoinformatics profiling of ionic liquids--uncovering structure-cytotoxicity relationships with network-like similarity graphs.

Ionic liquids (ILs) constitute one of the hottest areas in chemistry since they have become increasingly popular as reaction and extraction media. Their almost limitless structural possibilities, as opposed to limited structural variations within molecular solvents, make ILs "designer solvents." They also have been widely promoted as "green solvents" although their claimed relative nontoxicity has been frequently questioned. The Thinking in Structure-Activity Relationships (T-SAR) approach has proved to be an efficient method to gather relevant toxicological information of analog series of ILs. However, when data sets significantly grow in size and structural diversity, the use of computational models becomes essential. We provided such a computational solution in a previous work by introducing a reliable, predictive, simple, and chemically interpretable Classification and Regression Tree (CART) classifier enabling the prioritization of ILs with a favorable cytotoxicity profile. Even so, an efficient and exhaustive mining of SAR information goes beyond analog compound series and the applicability domain of quantitative SAR modeling. So, we decided to complement our previous findings based on the use of the CART classifier by applying the network-like similarity graph (NSG) approach to the mining of relevant structure-cytotoxicity relationship (SCR) trends. Finally, the SCR information concurrently gathered by both, quantitative (CART classifier) and qualitative (NSG) approaches was used to design a focused combinatorial library enriched with potentially safe ILs.

Chemoinformatics profiling of ionic liquids--automatic and chemically interpretable cytotoxicity profiling, virtual screening, and cytotoxicophore identification.

Ionic liquids (ILs) possess a unique physicochemical profile providing a wide range of applications. Their almost limitless structural possibilities allow the design of task-specific ILs. However, their "greenness," specifically their claimed relative nontoxicity has been frequently questioned, hindering their REACH registration processes and, so, their final application. Because the vast majority of ILs is yet to be synthesized, the development of chemoinformatics tools efficiently profiling their hazardous potential becomes essential. In this work, we introduce a reliable, predictive, simple, and chemically interpretable Classification and Regression Trees (CART) classifier, enabling the prioritization of ILs with a favorable cytotoxicity profile. Besides a good predictive capability (81% or 75% or 83% of accuracy or sensitivity or specificity in an external evaluation set), the other salient feature of the proposed cytotoxicity CART classifier is their simplicity and transparent chemical interpretation based on structural molecular fragments. The essentials of the current structure-cytotoxicity relationships of ILs are faithfully reproduced by this model, supporting its biophysical relevance and the reliability of the resultant predictions. By inspecting the structure of the CART, several moieties that can be regarded as "cytotoxicophores" were identified and used to establish a set of SAR trends specifically aimed to prioritize low-cytotoxicity ILs. Finally, we demonstrated the suitability of the joint use of the CART classifier and a group fusion similarity search as a virtual screening strategy for the automatic prioritization of safe ILs disperse in a data set of ILs of moderate to very high cytotoxicity.

Evolutionary computation and QSAR research.

The successful high throughput screening of molecule libraries for a specific biological property is one of the main improvements in drug discovery. The virtual molecular filtering and screening relies greatly on quantitative structure-activity relationship (QSAR) analysis, a mathematical model that correlates the activity of a molecule with molecular descriptors. QSAR models have the potential to reduce the costly failure of drug candidates in advanced (clinical) stages by filtering combinatorial libraries, eliminating candidates with a predicted toxic effect and poor pharmacokinetic profiles, and reducing the number of experiments. To obtain a predictive and reliable QSAR model, scientists use methods from various fields such as molecular modeling, pattern recognition, machine learning or artificial intelligence. QSAR modeling relies on three main steps: molecular structure codification into molecular descriptors, selection of relevant variables in the context of the analyzed activity, and search of the optimal mathematical model that correlates the molecular descriptors with a specific activity. Since a variety of techniques from statistics and artificial intelligence can aid variable selection and model building steps, this review focuses on the evolutionary computation methods supporting these tasks. Thus, this review explains the basic of the genetic algorithms and genetic programming as evolutionary computation approaches, the selection methods for high-dimensional data in QSAR, the methods to build QSAR models, the current evolutionary feature selection methods and applications in QSAR and the future trend on the joint or multi-task feature selection methods.