During the early ADME profiling the development of simple, interpretable and reliable in silico tools is very important. In this study, rule-based and QSPR approaches were investigated using a large Caco-2 permeability database. Three permeability classes were determined: high (H), moderate (M) and low (L). The main physicochemical properties related with permeability were ranked as follows: Polar Surface Area (PSA)>Lipophilicity (logP/logD)>Molecular Weight (MW)>number of Hydrogen Bond donors and acceptors>Ionization State>number of Rotatable Bonds>number of Rings. The best rule, based on the combination of PSA-MW-logD (3PRule), was able to identify the H, M and L classes with accuracy of 72.2, 72.9 and 70.6 %, respectively. Subsequently, a consensus system based on three voting binary classification trees was constructed. It accurately predicted 78.4/76.1/79.1 % of H/M/L compounds on training and 78.6/71.1/77.6 % on test set. Finally, the 3PRule and multiclassifier were validated with 23 drugs in a Caco-2 assay. The rule is very useful to improve assay design and prioritize the high absorption candidates. Meanwhile the QSPR model exhibits appropriate classification performance. Due to the simplicity, easy interpretation and accuracy, the 3PRule and consensus model developed here can be used in early ADME profiling.

Human small molecule metabolites (the human metabolome) are a set of compounds that interact with at least one macromolecule in the biosphere. This study investigates the acid/base profile of the human metabolome, natural products and drugs, together with an analysis of their physicochemical properties. Ionisation constants (pK_a values) are estimated for each compound and the identity of the ionisable functional groups in each set is determined. The acid/base and physicochemical property profile of the lipid component of the metabolome differed considerably to the other datasets. In contrast, the acid/base properties of non-lipid metabolites were found to be similar to both drugs and natural products. While the non-lipid metabolites have lower average ClogP values and more hydrogen bond donors than the other datasets, the distribution of physicochemical property values overlapped considerably with the drug dataset. Considering also that the non-lipid metabolites are of biochemical interest, their characteristics have great potential to influence the selection of screening compounds for drug discovery.

Many consensus clustering methods have been applied in different areas such as pattern recognition, machine learning, information theory and bioinformatics. However, few methods have been used for chemical compounds clustering. In this paper, an information theory and voting based algorithm (Adaptive Cumulative Voting-based Aggregation Algorithm A-CVAA) was examined for combining multiple clusterings of chemical structures. The effectiveness of clusterings was evaluated based on the ability of the clustering method to separate active from inactive molecules in each cluster, and the results were compared with Ward’s method. The chemical dataset MDL Drug Data Report (MDDR) and the Maximum Unbiased Validation (MUV) dataset were used. Experiments suggest that the adaptive cumulative voting-based consensus method can improve the effectiveness of combining multiple clusterings of chemical structures.

███

In the present study 2D-QSAR analysis was combined with information on crystallographic data and molecular modeling, in order to investigate dual PPAR-α/γ activity for a data set of 71 compounds, compiled from literature. Using Multivariate Data Analysis, satisfactory PLS models were generated for each receptor subtype separately. The models were based on simple and easily interpretable drug-like and constitutional descriptors, while the inclusion of MOLCONN-Z descriptors in the initial pool of variables had no considerable impact in model predictivity. By simultaneous analysis of both types of activity, a consensus PLS model for dual PPAR-α/γ activity could be derived, displaying the molecular features, which may lead to a balanced activity. All models were validated by permutation tests, by dividing the data set into training and test sets, as well as by external validation using a blind test set. Detailed inspection of PPAR-α and PPAR-γ crystal structures and molecular simulation supported the differentiation of most important descriptors in the separate PLS models, e.g. the higher impact of lipophilicity and bulk descriptors in PPAR-α and PPAR-γ activity respectively, as well as the effect of specific structural descriptors. Molecular simulation provided also explanation for the behavior of certain outliers in the PLS models.

Activity landscapes provide a comprehensive description of structure-activity relationships (SARs). An information theoretic assessment of their features, namely, activity cliffs, similarity cliffs, smooth-SAR, and featureless regions, is presented based on the probability of occurrence of these features. It is shown that activity cliffs provide highly informative SARs compared to smooth-SAR regions, although the latter are the basis for most QSAR studies. This follows since small structural changes in the former are coupled with relatively large changes in activity, thus pinpointing specific structural features associated with the changes in activity. In contrast, Smooth-SAR regions are typically associated with relatively small changes in both structure and activity. Surprisingly, similarity cliffs, which occur when both compounds in a compound-pair have approximately equal activities but significantly different structures, are the most prevalent feature of activity landscapes. Hence, from an information theoretic point of view, they are the least informative landscape feature. Nevertheless, similarity cliffs do provide SAR information on potentially new active compound classes, and in that sense they are quite useful in drug discovery programs since they provide alternative possibilities should ADMET or other issues arise during the discovery and earlier preclinical development phases of drug research.

Advances in the high-throughput determination of functional modulators of major histocompatibility complex (MHC) and improved computational predictions of MHC ligands have rendered the rational design of immunomodulatory peptides feasible. Proteome-derived peptides and ‘reverse vaccinology’ by computational means will play a driving role in future vaccine design. Here we review the molecular mechanisms of the MHC mediated immune response, present the computational approaches that have emerged in this area of biotechnology, and provide an overview of publicly available computational resources for predicting and designing new peptidic MHC ligands.

In this article, we compare the performance of 19 cluster validity indices, in identifying some possible genes mediating certain cancers, based on gene expression data. For the purpose of this comparison, we have developed a method. The proposed method involves cluster generation, selection of the best k-value or c-values, cluster identification, identifying the altered gene cluster, scoring an altered gene cluster and determining the best k-value or c-value exploring through biological repositories. The effectiveness of the method has been demonstrated on three gene expression data sets dealing with human lung cancer, colon cancer, and leukemia. Here, we have used three clustering algorithms, i.e., k-means, PAM and fuzzy c-means. We have used biochemical pathways related to these cancers and p-value statistics for validating the study.

Little published literature exists on the 881 bit structural keys used by PubChem for categorizing and comparing the compounds present in its database. We characterized these structural keys by examining their frequencies of occurrence within the PubChem compound database. In addition, bit dependencies, defined as the universal presence of a bit given the presence of another, were determined. We show that the vast majority of bits are rarely set and that substantial numbers of dependencies exist. A comparison of similarity searches with five United States Food and Drug Administration approved drugs as reference compounds using the full structural keys versus a variant in which all dependent bits were removed was performed using the Tanimoto coefficient. These bit dependencies not only affect similarity scores, but also alter the compounds returned in similarity searching. Judicious selection of bits is needed to maintain sufficient ability to differentiate related compounds.

Here we applied a novel method1a to predict pK_a values of the guanidine functional group, which is a notoriously difficult. This method, which was developed in our lab, uses only one ab initio bond length obtained at a low level of theory. The method is shown to work for drug molecules, delivers prediction errors of less than 0.5 log units, successfully treats tautomerisation in close relation with experiment, and demonstrates strong correlations with only a few data points. The high structural content of the ab initio bond length makes a given data set essentially divide itself into high correlation subsets. One then observes that molecules within a subset possess a common substructure. Each high correlation subset exists in its own region of chemical space. The high correlation subset method is explored with respect to this position in chemical space, in particular tautomerisation. The proposed method is able to distinguish between different tautomeric forms and the preferred tautomeric form emerges naturally, in agreement with experiment.

The enzyme aspartate β-semialdehyde dehydrogenase (ASADH) plays a vital role in biosynthesis of essential amino acids and several important metabolites in microbes and some higher plants. So this key enzyme can be targeted selectively in these microbes to exhibit anti-bacterial and fungicidal effects. In this work, molecular modeling and comparative active site binding mode studies were performed for understanding the mode of action, in silico insight into the 3D structure, enzyme-substrate interactions with natural substrate in this homologous enzyme family. During comparative sequence analysis, high diversity was found in the sequences of different ASADHs and exhibited the same key binding interactions with the substrate. Both, the functional carboxylic and the phosphate group of the substrate are engaged in a bidentate interaction with the guanidinium N atom of two key arginyl active site residues of ASADHs. These structural and active site binding mode characterization studies can further be used for designing the more potent and selective substrate analogues inhibitors against ASADH family.

β-Secretase (BACE1) is an aspartate protease involved in the production of amyloid-β a major peptide responsible for the pathogenesis of Alzheimer’s disease. Given its role in the formation of amyloids leading to Alzheimer’s disease, it has been a major therapeutic target for intervention and has been a challenge in the past and the progress has been very slow. More than hundred crystal structures with inhibitors are available in the protein data bank. Many strategies for drug design have been employed in the design of numerous diverse ligands for this target and many have failed due to undesirable drug properties primarily the inability to cross the blood-brain barrier. In the present work we attempted to consider multiple crystal structures with bound inhibitors showing affinity in the range of 2–210 nM efficacy and optimize the pharmacophoric requirement based on the energy involved in binding termed as e-pharmacophore mapping. A high throughput screening combined with molecular docking, ADMET predictions, logP values and in vitro assay led to the identification of 7 potential compounds showing inhibition at 10µM which could be further developed as novel inhibitors for β-secretase.