Bioisosteric approach in designing new monastrol derivatives: An investigation on their ADMET prediction using in silico derived parameters

Medicinal chemists are facing an increasing challenge to deliver safer and more effective medicines. An appropriate balance between drug-like properties such as solubility, permeability, metabolic stability, efficacy and toxicity is one of the most challenging problems during lead optimization of a potential drug candidate. Insoluble and impermeable compounds can result in erroneous biological data and unreliable SAR in enzyme and cell-based assays. The weak inhibitory activity and non-drug-like properties of monastrol, the first small mitotic kinesin Eg5 inhibitor, has hampered its further development. In this investigation, a bioisosteric approach was applied that resulted in the replacement of C-5 carbonyl of monastrol with thio-carbonyl. Further lead optimization of drug-like properties was evaluated through in silico predictions by using ADMET predictor software. This minor structural modification resulted in upgraded human effective jejunal permeability (Peff) and improved permeability in Madin–Darby canine kidney (MDCK) cells. Furthermore, C-5 thiocarbonyl analogue of monastrol (named as Special-2) was found safe to administer orally with no phospholipidosis toxicity, no raised levels of serum glutamate oxaloacetate transaminase (SGOT) and no potential towards cardiotoxicity. Molecular docking study was also carried out to understand the binding modes of these compounds. The docking study showed high binding affinity of the designed compounds against KSP. Hence a combination of in silico ADMET studies and molecular docking can help to improve prediction success and these compounds might be act as potential candidate for KSP inhibition.     

Modelling in vitro hepatotoxicity using molecular interaction fields and SIMCA

There is currently a great deal of interest in creating computational tools for predicting the pharmacological properties of drug development candidates, ranging from physicochemical properties such as pK_a and solubility to more complex biological properties such as oral bioavailability and toxicity. The limiting factor in many cases is a shortage of good data from which to construct training sets. In other cases, large amounts of data are available, but they use surrogate end-points or are comprised of compounds very different from those usually encountered in drug discovery and development. In such cases large training sets and global models are not necessarily better than local models based on smaller data sets. Such considerations make it as important to examine the available data carefully so as to avoid over-interpretation of the models obtained as it is to minimise errors in prediction per se. The kinds of complications likely to be encountered for in vitro hepatotoxicity modelling are discussed in general terms and illustrated in particular by SIMCA analysis of data obtained from assays of cultured hepatocytes for a large, structurally diverse data set and a smaller, much more focussed one.     

Investigation of the influence of molecular topology on ligand binding

Molecular topology class has previously been put forward as a new concept of describing compound quality and it has been shown that compared to general bioactive compounds, drugs is more similar to natural products and human metabolites in terms of molecular topology class distribution, in which they are enriched with compounds having only one ring system. To further understand how the molecular topology is influencing the drug discovery process, we have investigated the compound potency of different molecular topologies in published chemical patents. Our study shows that the potency for compounds having one ring system is higher compared to compounds that have more than one ring system. Compounds with one ring system are significantly less lipophilic and smaller compared to compounds with several ring systems. Further the influence of the molecular topology on ligand efficiency (LE), ligand lipophilic efficiency (LLE) and ligand-efficiency-dependent lipophilicity (LELP) was also analyzed and it was found that on average compounds with fewer ring systems and in particular compounds with only one ring system show consistently better LE, LLE and LELP. The results suggest that compounds with fewer ring systems and in particular compounds with only one ring system have good properties and that they might be useful starting point for drug discovery projects.     

SAHA-based novel HDAC inhibitor design by core hopping method

The catalytic activity of the histone deacetylase (HDAC) is directly relevant to the pathogenesis of cancer, and HDAC inhibitors represented a promising strategy for cancer therapy. SAHA (suberoanilide hydroxamic acid), an effective HDAC inhibitor, is an anti-cancer agent against T-cell lymphoma. However, SAHA has adverse effects such as poor pharmacokinetic properties and severe toxicities in clinical use. In order to identify better HDAC inhibitors, a compound database was established by core hopping of SAHA, which was then docked into HDAC-8 (PDB ID: 1T69) active site to select a number of candidates with higher docking score and better interaction with catalytic zinc ion. Further ADMET prediction was done to give ten compounds. Molecular dynamics simulation of the representative compound 101 was performed to study the stability of HDAC8-inhibitor system. This work provided an approach to design novel high-efficiency HDAC inhibitors with better ADMET properties.     

Identification of potential inhibitors against the Zika virus using consensus scoring

The Zika virus (ZIKV) is a life threatening pathogen of zoonotic importance with prevalence in some parts of Africa and America. Unfortunately, there is yet to be a single approved vaccine or antiviral drug to treat the diseases and deformations being caused by the Zika virus infection. In this study, about 36 million compounds from MCULE database were virtually screened against a real matured ZIKV protein using a consensus scoring method to get improved hit rates. The consensus scoring method combined the result from the 25 top ranked molecules from both MCULE and Drug Score eXtended (DSX) docking programs which led to the selection of two hit compounds. The inhibition constant (Ki) values of 0.08 and 0.30 μm were obtained for the two selected compounds MCULE-8830369631-0-1 and MCULE-9236850811-0-1 respectively, to remark them as hit compounds. The molecular interactions of the two selected hit compounds with the amino acids (ALA 48, ILE 49, ILE 468 and LEU 472) present in the ZIKV protein indicated that they both have similar binding modes. The result of the computationally predicted physicochemical properties including ADMET for the selected compounds showed their great potential in becoming lead compounds upon optimization and thus could be used in treating the Zika virus diseases.     

一种融合胶囊网络的分类方法

目前的ADMET分类方法在对具有多特征性和特征关联性的化合物数据进行ADMET分类时存在不足。而且,对ADMET分类结果不具备可解释性。针对上述问题,提出一种融合胶囊网络的分类模型（CapsMC）。CapsMC模型首先提出一种feature-to-image图像转换算法。使用该算法将特征之间的关联关系和依赖关系作为考量纳入到分类依据中,实现特征的多层次提取。其次,探索胶囊网络的高级应用,提出一种认知推理机制。使用该机制对特征进行认知推理,实现ADMET的可解释性分类。模型在五种ADMET数据集上的实验结果表明,CapsMC模型可以高效实现ADMET的可解释性分类。     

Genetic programming for computational pharmacokinetics in drug discovery and development

The success of a drug treatment is strongly correlated with the ability of a molecule to reach its target in the patient’s organism without inducing toxic effects. Moreover the reduction of cost and time associated with drug discovery and development is becoming a crucial requirement for pharmaceutical industry. Therefore computational methods allowing reliable predictions of newly synthesized compounds properties are of outmost relevance. In this paper we discuss the role of genetic programming in predictive pharmacokinetics, considering the estimation of adsorption, distribution, metabolism, excretion and toxicity processes (ADMET) that a drug undergoes into the patient’s organism. We compare genetic programming with other well known machine learning techniques according to their ability to predict oral bioavailability (%F), median oral lethal dose (LD50) and plasma-protein binding levels (%PPB). Since these parameters respectively characterize the percentage of initial drug dose that effectively reaches the systemic blood circulation, the harmful effects and the distribution into the organism of a drug, they are essential for the selection of potentially good molecules. Our results suggest that genetic programming is a valuable technique for predicting pharmacokinetics parameters, both from the point of view of the accuracy and of the generalization ability.

ChemoGraph: Interactive Visual Exploration of the Chemical Space

Exploratory analysis of the chemical space is an important task in the field of cheminformatics. For example, in drug discovery research, chemists investigate sets of thousands of chemical compounds in order to identify novel yet structurally similar synthetic compounds to replace natural products. Manually exploring the chemical space inhabited by all possible molecules and chemical compounds is impractical, and therefore presents a challenge. To fill this gap, we present ChemoGraph, a novel visual analytics technique for interactively exploring related chemicals. In ChemoGraph, we formalize a chemical space as a hypergraph and apply novel machine learning models to compute related chemical compounds. It uses a database to find related compounds from a known space and a machine learning model to generate new ones, which helps enlarge the known space. Moreover, ChemoGraph highlights interactive features that support users in viewing, comparing, and organizing computationally identified related chemicals. With a drug discovery usage scenario and initial expert feedback from a case study, we demonstrate the usefulness of ChemoGraph.     

Associative Neural Network

An associative neural network (ASNN) is a combination of an ensemble of the feed-forward neural networks and the K-nearest neighbor technique. The introduced network uses correlation between ensemble responses as a measure of distance among the analyzed cases for the nearest neighbor technique and provides an improved prediction by the bias correction of the neural network ensemble both for function approximation and classification. Actually, the proposed method corrects a bias of a global model for a considered data case by analyzing the biases of its nearest neighbors determined in the space of calculated models. An associative neural network has a memory that can coincide with the training set. If new data become available the network can provide a reasonable approximation of such data without a need to retrain the neural network ensemble. Applications of ASNN for prediction of lipophilicity of chemical compounds and classification of UCI letter and satellite data set are presented. The developed algorithm is available on-line at http://www.virtuallaboratory.org/lab/asnn.     

Computing as compression: SP20

SP20 is a software simulation of a ‘new generation’ computing system, based on the conjecture that computing may usefully be understood as information compression by pattern matching, unification and metrics-guided search.¹⁹⁾ The design of the system aims to exploit the potential of these processes as fully as possible to integrate and rationalise diverse kinds of computing and to achieve more flexibility and ‘intelligence’ than conventional computers. The organisation of SP20 is described, highlighting advances compared with previous models. The main advances are: a much more efficient search method which is scaleable to large data sets; an improved ability to find alternative answers to problems; and an ability to find patterns which are discontinuous or fragmented as well as coherent patterns of contiguous symbols. Analytic and empirical evidence is presented confirming the computational properties of the model. In a serial processing environment, its time complexity is O(N ²), whereN is the number of symbols processed. In a high-parallel environment, the time complexity of this model should approach O(N). The space complexity of the model in serial or parallel environments appears to be O(N). The capabilities and shortcomings of SP20 are described in different areas of computing: best-match information retrieval, pattern recognition and de-referencing of identifiers; unsupervised inductive learning of grammars, object-oriented structures and rules; execution of functions; deductive and probabilistic inference; parsing; planning and problem solving. A selection of examples are presented, highlighting the new capabilities of the model. Weaknesses in the design of the model are summarised and planned future developments are described in outline.     

3D-QSAR studies of D3R antagonists and 5-HT1AR agonists

Combination of dopamine D₃ antagonism and serotonin 5-HT_1A agonism leads to an effective way to atypical antipsychotics. In this work, two predictive 3D-QSAR models were bulit for D₃R antagonists and 5-HT_1AR agonists, respectively. Based on the steric and electrostatic information of contour maps, four compounds with improved predicted activities were newly designed. In addition, molecular docking and ADMET properties suggested that designed molecules had strong interactions with receptors and low hepatotoxicity. This work sheds light on the design of bifunctional novel antipsychotic drugs for D₃R antagonists and 5HT_1AR agonists.     

Targeting the dengue β-OG with serotype-specific alkaloid virtual leads

The dengue envelope β-OG pocket is a crucial hinge for mediating virus-host fusion via conformational changes in the envelope to the fusion-competent form. The β-OG pocket is a small molecule target site for inhibition of virus-host fusion. As of date, the only structure of the β-OG pocket known is of serotype 2. Studies of β-OG inhibition by small molecules primarily target viral serotype 2. Envelope and β-OG sequence alignments, reveal dissimilarities across serotypes. In light of protein sequence-structure-function correlation, sequence variations suggest serotypic variations in β-OG druggability. This, together with the fact that dengue viral proteins do have serotype-specific variations of structure and function, lead to the study of the serotype-specificity of the dengue β-OG ligand binding behaviour. β-OG druggability was compared using comparative models of envelope proteins containing the β-OG pocket in four serotypes of the dengue virus. β-OG ligand binding was found to vary with respect to hydrophobicity, hydrophilicity, hydrogen bonding, van der Waals interactions with ligands and tightness of the binding site. The study also reports serotype-specific virtual leads identified from a library of 9175 alkaloids, using a consensus docking and scoring approach. The docking algorithms of Glide SP and XP, together with the Lamarckian genetic algorithm were employed for consensus docking. For consensus scoring, the Glide empirical score was employed along with the scoring function of AutoDock. A multi-dimensional lead optimisation approach was performed for optimising affinity, ligand efficiency, lipophilic ligand efficiency, ADMET and molecular torsional strains. The study proposes the serotype-specific inhibition of the β-OG for an effective inhibition of virus-host fusion, in contrast to a pan inhibitor.