Current Computer - Aided Drug Design - Volume 18, Issue 2, 2022

Background: The manual segmentation of cellular structures on Z-stack microscopic images is time-consuming and often inaccurate, highlighting the need to develop auto-segmentation tools to facilitate this process. Objective: This study aimed to compare the performance of three different machine learning architectures, including random forest (RF), AdaBoost, and multi-layer perceptron (MLP), for the autosegmentation of nuclei in proliferating cervical cancer cells on Z-Stack cellular microscopy proliferation images provided by the HCS Pharma. The impact of using post-processing techniques, such as the StarDist plugin and majority voting, was also evaluated. Methods: The RF, AdaBoost, and MLP algorithms were used to auto-segment the nuclei of cervical cancer cells on microscopic images at different Z-stack positions. Post-processing techniques were then applied to each algorithm. The performance of all algorithms was compared by an expert to globally generated ground truth by calculating the accuracy detection rate, the Dice coefficient, and the Jaccard index. Results: RF achieved the best accuracy, followed by the AdaBoost and then the MLP. All algorithms achieved good pixel classifications except in regions whereby the nuclei overlapped. The majority voting and StarDist plugin improved the accuracy of the segmentation but did not resolve the nuclei overlap issue. The Z-Stack analysis revealed similar segmentation results to the Z-stack layer used to train the image. However, a worse performance was noted for segmentations performed on different Z-stack positions, which were not used to train the algorithms. Conclusion: All machine learning architectures provided a good segmentation of nuclei in cervical cancer cells but did not resolve the problem of overlapping nuclei and Z-stack segmentation. Further research should therefore evaluate the combined segmentation techniques and deep learning architectures to resolve these issues.

Background: The fragment-to-fragment approach for the estimation of the biological affinity of the pharmacophores with biologically active molecules has been proposed. It is the next step in the elaboration of molecular docking and using the quantum-chemical methods for the complex modeling of pharmacophores with biomolecule fragments. Methods: The parameter φ 0 was used to estimate the contribution of π-electron interactions in biological affinity. It is directly related to the position of the frontier levels and reflects the donor-acceptor properties of the pharmacophores and stabilization energy of the [Pharm158;‰BioM] complex Results: By using quantum-chemical calculations, it was found that the stacking interaction of oxazoles with phenylalanine is 7-11 kcal/mol, while the energy of hydrogen bonding of oxazoles with the amino group of lysine is 5-9 kcal/mol. The fragment-to-fragment approach can be applied for the investigation of the dependence of biological affinity on the electronic structure of pharmacophores.c Conclusion: The founded quantum-chemical regularities are confirmed with the structure-activity relationships of substituted oxazoles.

Background: Smokeless tobacco (SLT) is traditionally used in Middle East countries. The several toxic constituents with potential carcinogenicity make it a serious human health risk. Literature regarding their effects on cardiac and cancer disease is lacking in Saudi Arabia. Objective: This study was conducted to investigate the adverse effect of 11 different samples of widely used SLT varieties from the Tabuk region - Saudi Arabia, on Nitric Oxide (NO) level and their potential risk on cardiovascular health, etiology and/or progression of cancers. Methods: Samples were collected from Tabuk, KSA and analyzed by the GC-MS technique. Nitric oxide inhibition was performed using J774.2 macrophages by the Griess method. The retrieved crystallized structure of human inducible nitric oxide synthase (iNOS) from Brookhaven Protein Data Bank Repository PDB I.D: 3E7G with 2.20Å resolution was further prepared by structure using the MOE.2019 tool. The compounds abstracted from 11 different Shammah varieties were sketched by the MOE-Builder tool. Minimization for both receptor and compounds was performed via AMBER99 and MMFF99X force field implemented in MOE. Results: Nine samples (4 - 11) showed a potent suppressive effect on NO production with IC50 values ranging between (16.9-20.4 μg/mL), respectively. The samples (1 & 2) exhibited a moderate level of inhibition with IC50 ranging between 33.2 and 57.4 μg/mL, respectively. Interestingly, sample 4 consisting of compounds (13-15, 19-26, 28) that mostly belongs to the group fatty acid ester and phthalic acid ester showed the most potent suppressive effect. Molecular docking results revealed that the current local SLT constituents presented noticeable potency in different extract samples. Conclusion: Variable suppressive effects on NO were detected in the current SLT samples, where sample 4 was the most potent among all. The extract of the latter exhibited molecular interaction with the first shell amino acid residues of Inducible nitric oxide synthase (iNOS), which may anchor the plasticity and selectivity of the compounds present in it. The samples (4 -11) showed a potent inhibitory effect on the NO, where compound 26 (Phthalic acid ester) is common, and its adequate concentration may account for augmented biological activity. These results may effectively highlight their adverse effects on cardiovascular health and etiology and/or progression of cancer and may help in strengthening the social and governmental efforts in minimizing the use of these substances.

Background: Alzheimer’s disease (AD) is a neurodegenerative condition and the most common type of dementia among the elderly. The enzymes acetylcholinesterase (AChE) and nitric oxide synthase (NOS) have a pivotal role in the pathophysiology of this disease. Objective: This study aimed to select medicinal plant-derived molecules with reported inhibition of AChE and design optimized molecules that could inhibit not only AChE, but also NOS, potentially increasing its efficacy against AD. Methods: 24 compounds were selected from the literature based on their known AChE inhibitory activity. Then, we performed molecular orbital calculations, maps of electrostatic potential, molecular docking study, identification of the pharmacophoric pattern, evaluation of pharmacokinetic and toxicological properties of these molecules. Next, ten analogs were generated for each molecule to optimize their effect where the best molecules of natural products had failed. Results: The most relevant correlation was between HOMO and GAP in the correlation matrix of the molecules’ descriptors. The pharmacophoric group’s derivation found the following pharmacophoric features: two hydrogen bond acceptors and one aromatic ring. The studied molecules interacted with the active site of AChE through hydrophobic and hydrogen bonds and with NOS through hydrogen interactions only but in a meaningful manner. In the pharmacokinetic and toxicological prediction, the compounds showed satisfactory results. Conclusion: The design of natural products analogs demonstrated good affinities with the pharmacological targets AChE and NOS, with satisfactory pharmacokinetics and toxicology profiles. Thus, the results could identify promising molecules for treating Alzheimer’s disease.

Background: The vast diversity of peptide sequences may hinder the effectiveness of screening for potential peptide therapeutics as if searching for a needle in a haystack. This study aims to develop a new self-evolving peptide algorithm (SEPA), for easy virtual screening of small linear peptides (three to six amino acids) as potential therapeutic agents with the collaborative use of freely available software that can be run on any operating system equipped with a Bash scripting terminal. Mitogen-inducible Gene 6 (Mig6) protein, a cytoplasmic protein responsible for inhibition and regulation of epidermal growth factor receptor tyrosine kinase, was used to demonstrate the algorithm. Objective: The objective is to propose a new method to discover potential novel peptide inhibitors via an automated peptide generation, docking and post-docking analysis algorithm that ranks short peptides by using essential hydrogen bond interaction between peptides and the target receptor. Methods: A library of dockable dipeptides were first created using PyMOL, Open Babel and AutoDockTools, and docked into the target receptor using AutoDock Vina, automatically via a Bash script. The docked peptides were then ranked by hydrogen bond interaction-based thorough interaction analysis, where the top-ranked peptides were then elongated, docked, and ranked again. The process repeats until the user-defined peptide length is achieved. Results: In the tested example, SEPA bash script was able to identify the tripeptide YYH ranked within top 20 based on the essential hydrogen bond interaction towards the essential amino acid residue ASP837 in the EGFR-TK receptor. Conclusion: SEPA could be an alternative approach for the virtual screening of peptide sequences against drug targets.