Publications

Abstract Nanotechnology has led to the development of new nanomaterials with unique properties and a wide variety of applications. In the present study, we focused on the cellular uptake of a group of nanoparticles with a single metal core by pancreatic cancer cells, which has been studied by Yap et al. (Rsc Advances, 2012, 2 (2):8489-8496) using classification models. In this work, the development of a further Quantitative Nanostructure Activity Relationship (QNAR) model was performed by linear multiple linear regression (MLR) and nonlinear artificial neural network (ANN) techniques to accurately predict the cellular uptake values of these compounds by dividing them into three groups. Judging from the attained statistical results, our derived QNAR models have an acceptable overall accuracy and robustness, as well as good predictivity on the external data sets. Moreover, the results of this study provide some insights on how engineered nanomaterial features influence cellular responses and thereby outline possible approaches for developing and applying predictive computational models for biological responses caused by exposure to nanomaterials.

Abstract 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.

Abstract Adenosine receptors (ARs) are signaling molecules ubiquitously expressed in a wide variety of tissues in the human body. ARs mediate physiological functions by interacting with four subtypes of G-protein-coupled receptors, namely A(1), A(2A), A(2B) and A(3). The A(3) AR, probably the most studied subtype, is also ubiquitously expressed, with high levels in peripheral organs and low levels in the brain. This type of AR is involved in a variety of important pathophysiological processes, ranging from modulation of cerebral and cardiac ischemic damage to regulation of immunosuppression and inflammation. Consequently, the development of potent and selective A(3) AR ligands as promising therapeutic options for a variety of diseases has been a prime subject of medicinal chemistry research for more than two decades. Among the plethora of approaches applied quantitative structure activity relationships (QSAR) stands out for being largely employed due to their potential to increase the efficiency at initial stages of the drug discovery process. So, we provide a review of the main QSAR studies devoted to the design, discovery and development of agonist and antagonist A(3) adenosine receptor ligands. Common pitfalls of these QSAR applications and the current trends in this area are also analyzed.

Abstract Today, emerging and increasing resistance to antibiotics has become a threat to public health worldwide. Antimicrobial peptides own unique action mechanisms making peptide antibiotics an attractive therapeutic option against resistant bacteria. However, their high haemolytic activity lacks the selectivity required for a human antibiotic. Therefore, additional efforts are needed to develop new antimicrobial peptides that possess greater selectivity for bacterial cells over erythrocytes. In this article, we introduce a chemoinformatics approach to simultaneously deal with these two conflicting properties consisting on a multi-criteria virtual screening strategy based on the use of a desirability-based multi-criteria classifier combined with similarity and chemometrics concepts. Here we propose a new quantitative feature encoding information related to the desirability, the degree of credibility ascribed to this desirability and the similarity of a candidate to a highly desirable query, which can be used as ranking criterion in a virtual screening campaign, the Desirability-Credibility-Similarity (DCS) Score. The enrichment ability of a multi-criteria virtual screening strategy based on the use of the DCS Score it is also assessed and compared to other virtual screening options. The results obtained evidenced that the use of the DCS score seems to be an efficient virtual screening strategy rendering promising overall and initial enrichment performance. Specifically, by using the DCS score it was possible to rank a selective antibacterial peptidomimetic earlier than a biologically inactive or non selective antibacterial peptidomimetic with a probability of ca. 0.9.

Abstract 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.

Abstract The adjustment of multiple criteria in hit-to-lead identification and lead optimization is a major advance in drug discovery. Thus, the development of approaches able to handle additional criteria for the early simultaneous treatment of the most important properties determining the pharmaceutical profile of a drug candidate is an emergent issue in this area. In this paper, we review a desirability-based multi-objective QSAR method allowing the joint handling of multiple properties of interest in drug discovery: the MOOP-DESIRE methodology. This methodology adapts desirability theory concepts allowing the holistic modeling of the many and conflicting biological properties determining the therapeutic utility of a drug candidate. Here we survey their suitability for key tasks involving the use of chemoinformatics methods in medicinal chemistry and drug discovery.

Abstract Blood Serum Proteome-Mass Spectra (SP-MS) may allow detecting Proteome-Early Drug Induced Cardiac Toxicity Relationships (called here Pro-EDICToRs).However, due to the thousands of proteins in the SP, a more realistic alternative representsthe identification of general Pro-EDICToRs patterns instead of a single protein marker. Inthis sense, we introduced a novel Cartesian 2D spectrum graph for SP-MS. Next, wecalculated the graph node-overlapping parameters (nopk) to numerically characterize SPMSby using them as inputs for a Quantitative Proteome-Toxicity Relationship (QPTR)classifier for Pro-EDICToRs with accuracy higher than 80%. This QPTR approach is theresult of adapting the classic blood proteome Quantitative Property-Structure Relationshipmodels (QSPR) used in Chemometrics to low-mass molecules study. Principal ComponentAnalysis (PCA) on the QPTR nopk values explains with one factor (F1) the 82.7% ofvariance. These nopk values were used to construct for the first time a Pro-EDICToRsComplex Network having samples as nodes linked by similarity between two samplesedges. We compared the topology of two sub-networks for the cardiac toxicity and controlsamples and found extreme relative differences for the re-linking (P) and Zagreb (M2)indices (9.5 and 54.2 % respectively) out of 11 parameters. We also compared the subnetworkswith the well-known ideal random networks including Barabasi-Albert,Kleinberg Small World, Erdos-Renyi, and Epsstein Power Law models. Finally, weproposed Partial Order (PO) schemes of the 115 samples based on LDA-probabilities, F1-scores and/or network node degrees. PCA-CN and LDA-PCA based POs with Tanimoto's.

Abstract In the present work, Dempster-Shafer Theory (DST) was employed for the implementation of a combined strategy for classification and/or virtual screening of potential anticoccidial drug candidates, based on the combination of the information provided by multiple QSAR models which are derived from different molecular structure representations. The application of such a strategy lead to a classification performance superior to the individual use of QSAR models, achieving accuracy/sensibility/specificity values over 94%/86%/96% and 86%/75%/89% on training and predicting series, respectively. Parallely, the application of such a strategy lead to values of enrichment metrics significantly superiors to the individual use of QSAR models as virtual screening tools. All these results suggest that the use of DST as the theoretical probabilistic base for the implementation of a combined classification and/or virtual screening strategy can be efficiently employed on the process of discovery and development of novel potential anticoccidial candidates, contributing in this way to overcome the emergence of resistance to current therapies.

Abstract Today, emerging and increasing resistance to antibiotics has become a threat to public health worldwide. Antimicrobial peptides have unique action mechanisms making them an attractive therapeutic prospect to be applied against resistant bacteria. However, the major drawback is related with their high hemolytic activity which cancels out the safety requirements for a human antibiotic. Therefore, additional efforts are needed to develop new antimicrobial peptides that possess a greater potency for bacterial cells and less or no toxicity over erythrocytes. In this paper, we introduce a practical approach to simultaneously deal with these two conflicting properties. The convergence of machine learning techniques and desirability theory allowed us to derive a simple, predictive, and interpretable multicriteria classification rule for simultaneously handling the antibacterial and hemolytic properties of a set of cyclic beta-hairpin cationic peptidomimetics (C beta-HCPs). The multicriteria classification rule exhibited a prediction accuracy of about 80% on training and external validation sets. Results from an additional concordance test have shown an excellent agreement between the multicriteria classification rule predictions and the predictions from independent classifiers for complementary antibacterial and hemolytic activities, respectively, evidencing the reliability of the multicriteria classification rule. The rule was also consistent with the general mode of action of cationic peptides pointing out its biophysical relevance. We also propose a multicriteria virtual screening strategy based on the joint use of the multicriteria classification rule, desirability, similarity, and chemometrics concepts. The ability of such a virtual screening strategy to prioritize selective (nonhemolytic) antibacterial C beta-HCPs was assessed and challenged for their predictivity regarding the training, validation, and overall data. In doing so, we were able to rank a selective antibacterial C beta-HCP earlier than a biologically inactive or nonselective antibacterial C beta-HCP with a probability of ca. 0.9. Our results thus indicate that promising chemoinformatics tools were obtained by considering both the multicriteria classification rule and the virtual screening strategy, which could, for instance, be used to aid the discovery and development of potent and nontoxic antimicrobial peptides.

Abstract Desirability theory (DT) is a well-known multi-criteria decision-making approach. In this work, DT is employed as a prediction model (PM) interpretation tool to extract useful information on the desired trade-offs between binding and relative efficacy of N6-substituted-4'-thioadenosines A(3) adenosine receptor (A(3)AR) agonists. At the same time, it was shown the usefulness of a parallel but independent approach providing a feedback on the reliability of the combination of properties predicted as a unique desirability value. The appliance of belief theory allowed the quantification of the reliability of the predicted desirability of a compound according to two inverse and independent but complementary prediction approaches. This information is proven to be useful as a ranking criterion in a ligand-based virtual screening study. The development of a linear PM of the A(3)AR agonists overall desirability allows finding significant clues based on simple molecular descriptors. The model suggests a relevant role of the type of substituent on the N6 position of the adenine ring that in general contribute to reduce the flexibility and hydrophobicity of the lead compound. The mapping of the desirability function derived of the PM offers specific information such as the shape and optimal size of the N6 substituent. The model herein developed allows a simultaneous analysis of both binding and relative efficacy profiles of A(3)AR agonists. The information retrieved guides the theoretical design and assembling of a combinatorial library suitable for filtering new N6-substituted-4'-thioadenosines A(3)AR agonist candidates with simultaneously improved binding and relative efficacy profiles. The utility of the desirability/belief-based proposed virtual screening strategy was deduced from our training set. Based on the overall results, it is possible to assert that the combined use of desirability and belief theories in computational medicinal chemistry research can aid the discovery of A(3)AR agonist candidates with favorable balance between binding and relative efficacy profiles.