China
Africa
Explore chapters and articles related to this topic
Perception, Planning, and Scoping, Problem Formulation, and Hazard Identification
Published in Ted W. Simon, Environmental Risk Assessment, 2019
Read-across is a technique that uses toxicity data from chemicals with structures similar to the one under consideration.184 These structurally similar chemicals are then used to restrict the domain of application of QSAR results.185 An associated area of interest is the grouping of chemicals and the identification of analogs. Analog identification may be based on presence within a congeneric series, similar functional groups, overall chemical similarity, mechanism of action, or 3-D structure.186–189 Graphic and statistical methods for displaying chemical similarity via network mapping are being used to communicate the correspondence (or lack thereof) between the shared structure of chemicals and biological activity.190–194
Aircraft Decontamination and Mitigation
Published in Brian J. Lukey, James A. Romano, Salem Harry, Chemical Warfare Agents, 2019
William T. Greer Jr., Angela M.G. Theys, William R. Davis, Kenneth J. Heater
The number and type of chemical simulants are as varied as the chemical agents themselves, and it is difficult to find simulants that mimic every aspect of agent behavior without having similar toxicity. Therefore, chemical simulants are tailored to the behavior and chemical characteristic being studied (Bartelt-Hunt et al., 2008). For example, where weathering and thermal properties are concerned, one of the most important characteristics of the simulant will likely be the vapor pressure or rate of evaporation. Simulants for this type of study may have very little chemical similarity to the agent in question, but the similarity in vapor pressure makes them a sound choice. However, if the study involves reacting or detoxifying the agent, it is more appropriate to choose a simulant that is chemically similar to the agent with regard to functionality, as opposed to having similar physical characteristics such as vapor pressure.
Emerging Recreational Psychotropics
Published in David J. George, Poisons, 2017
There is an understandable tendency to group the emerging psychoactive substances by their chemical similarity to other more well-known substances. For example, a substance may be a derivative or an analog of amphetamine. This might imply to users that the new substance is merely a stronger or more potent form of amphetamine. This is not generally the case. Small modifications in the chemical structure of a compound can produce significant changes in the spectrum of effects that the newer compound might produce. The newer compound might be more potent, but it might also have effects that are qualitatively much different from the parent compound. When chemical structures of psychoactive chemicals are presented side-by-side in a pictorial format, some may appear very similar to one another but the similarity usually ends there. Their effects can vary significantly.
The application of machine learning techniques to innovative antibacterial discovery and development
Published in Expert Opinion on Drug Discovery, 2020
Mateus Sá Magalhães Serafim, Thales Kronenberger, Patrícia Rufino Oliveira, Antti Poso, Káthia Maria Honório, Bruno Eduardo Fernandes Mota, Vinícius Gonçalves Maltarollo
The work from Maltarollo, 2019 [114] described the application of several machine learning techniques on a series containing 166 known inhibitors and non-inhibitors of S. aureus enoyl acyl carrier protein reductase (FabI). FabI is a key enzyme in fatty-acid metabolism and the molecular target of the antibiotic triclosan. SVM, MLP and RF methods generated the most predictive models with Mathews correlation coefficient (MCC) values for classifying the test set compounds higher than 0.75. Interestingly, a more recent virtual screening work employed three-dimensional chemical similarity statistically validated models with similar performance of previously described work (MCC values ranging from 0.73 to 0.79) [115]. In this research, four compounds with antibacterial activity against E. coli, S. aureus and MRSA were found using ligand-based virtual screening. The models were generated with the same dataset and predicted important features (functional groups at specific positions) similarly to the MLT models. This comparison indicates that MLTs could find similar and comparable patterns to other methods, in this case the 3D chemical similarity.
MS-275 Chemical Analogues Promote Hemoglobin Production and Erythroid Differentiation of K562 Cells
Published in Hemoglobin, 2019
Stella Voskou, Marios Phylactides, Andreas Afantitis, Georgia Melagraki, Andreas Tsoumanis, Panayotis A. Koutentis, Tina Mitsidi, Styliana I. Mirallai, Marina Kleanthous
The identification of chemical analogues of MS-275 was performed in three consecutive rounds of selection. For the first two rounds, molecular modeling was applied and the PubChem library was screened for potential pharmacological agents. PubChem contains more than 5 million compounds and is currently the largest publicly available chemical database. The Enalos Mold2 KNIME node [35,36] was used to compute the 42 description pharmacophores on which the chemical similarity metrics employed for the selection of the test chemicals were based. These 42 molecular quantum numbers reflect the subset of molecular descriptors deemed to be important for the compounds’ underlying biological activity. The descriptors were computed to account for chemical, physicochemical and electronic properties of molecules.
Advances in distributed computing with modern drug discovery
Published in Expert Opinion on Drug Discovery, 2019
Antonio Jesús Banegas-Luna, Baldomero Imbernón, Antonio Llanes Castro, Alfonso Pérez-Garrido, José Pedro Cerón-Carrasco, Sandra Gesing, Ivan Merelli, Daniele D’Agostino, Horacio Pérez-Sánchez
Many computational techniques are now available to study the molecular interactions relevant to drug discovery, such as virtual screening (VS) [3], which is used to simulate a large number of interactions between proteins (also known as receptors and/or enzymes) and small molecule drug candidates (ligands). Docking software is usually tested on protein families where there are the most crystal structures and, therefore, it is common practice to test many proteins in parallel [4–6]. Side-effects caused by off-target bindings should be avoided and, therefore, the most promising compounds are usually tested against many other proteins. The docking conformations that describe the interactions between each compound and the corresponding target are optimized through molecular dynamics (MD) simulations to relax the system and improve the accuracy with which the binding energy is calculated. MD is a physics-based simulation method in which Newton’s equations of motion are solved for each atom of the system considering all the forces involved in their interactions [7]. Depending on the number of atoms involved, it can be computationally demanding. Other cheaper techniques, such as chemical similarity and calculating the proximity matrix, are also used in the drug discovery process [8–10].