Tripeptides from Allium subhirsitum L. extracts: Pharmacokinetics properties, toxicity prediction and in silico study against SARS-CoV-2 enzymes and pro-inflammatory proteins

Developing new prophylactic and therapeutic agents with broad-spectrum antiviral activities is urgently needed to combat emerging human severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Since no available clinically antiviral drugs have been approved to eradicate COVID-19 as of the writing of this report, this study aimed to investigate bioactive short peptides from Allium subhirsutum L. (Hairy garlic) extracts identified through HR-LC/MS analysis that could potentially hinder the multiplication cycle of SARS-CoV-2 via molecular docking study. The obtained promising results showed that the peptides (Asn-Asn-Asn) possess the highest binding affinities of -8.4 kcal/mol against S protein, (His-Phe-Gln) of -9.8 kcal/mol and (Gln-His-Phe) of -9.7 kcal/mol towards hACE2, (Thr-Leu-Trp) of -10.3 kcal/mol and (Gln-Phe-Tyr) of -9.8 kcal/mol against furin. Additionally, the identified peptides show strong interactions with the targeted and pro-inflammatory ranging from -8.1 to -10.5 kcal/mol for NF−κB-inducing kinase (NIK), from -8.2 to -10 kcal/mol for phospholipase A2 (PLA2), from -8.0 to -10.7 kcal/mol for interleukin-1 receptor-associated kinase 4 (IRAK-4), and from -8.6 to -11.6 kcal/mol for the cyclooxygenase 2 (COX2) with Gln-Phe-Tyr model seems to be the most prominent. Results from pharmacophore, drug-likeness and ADMET prediction analyses clearly evidenced the usability of the peptides to be developed as an effective drug, beneficial for COVID-19 treatment.

Study of Molecular Docking, Molecular Dynamic and Toxicity Prediction of Several Quinoline Alkaloid Derivatives as a Bruton Tyrosine Kinase Inhibitor as Anti-Leukemia

Quinoline alkaloid and its derivatives play a vital role in the development of new therapeutic agents. Cinnoline structure has similarities with quinoline alkaloid compound and has the potential to inhibit Bruton’s Tyrosine Kinase (BTK) in leukemia treatment. This research aims to study the interaction of several quinoline alkaloids with BTK and to predict the toxicity to ensure their safety. This study was carried out using computational studies, including molecular docking, molecular dynamics simulation, and toxicity prediction, to assess the compound’s activity towards BTK and their toxicity. Molecular docking simulations showed that ten compounds (S1, S2, S4, S5, S8, S13, S14, S16, S17, and S20) had the best affinity to BTK. Molecular dynamics simulations to these ten compounds showed that only seven compounds (S1, S5, S8, S13, S16, S17, and S20) could stabilize the interaction towards BTK with RMSD and RMSF value of 0.5 ± 2 Å and 0.5 ± 6, 5 Å, respectively. Toxicity prediction results showed that these quinoline alkaloids had various toxicity characteristics, but most were not carcinogens and mutagens (S4, S5, S6, S7, S8, S10 S11, S12, S14, and S15). It can be concluded that Yukositrin (S8) has the most potential affinity towards BTK, which can be used as anti-leukemia with low toxicity. Keywords: anti-leukemia, Bruton Tyrosine Kinase, docking, MD, quinoline alkaloids

Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network

As safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are these experiments time consuming and costly, but experiments that involve animal testing are increasingly subject to ethical concerns. While traditional machine learning (ML) methods have been used in the field with some success, the limited availability of annotated toxicity data is the major hurdle for further improving model performance. Inspired by the success of semi-supervised learning (SSL) algorithms, we propose a Graph Convolution Neural Network (GCN) to predict chemical toxicity and trained the network by the Mean Teacher (MT) SSL algorithm. Using the Tox21 data, our optimal SSL-GCN models for predicting the twelve toxicological endpoints achieve an average ROC-AUC score of 0.757 in the test set, which is a 6% improvement over GCN models trained by supervised learning and conventional ML methods. Our SSL-GCN models also exhibit superior performance when compared to models constructed using the built-in DeepChem ML methods. This study demonstrates that SSL can increase the prediction power of models by learning from unannotated data. The optimal unannotated to annotated data ratio ranges between 1:1 and 4:1. This study demonstrates the success of SSL in chemical toxicity prediction; the same technique is expected to be beneficial to other chemical property prediction tasks by utilizing existing large chemical databases. Our optimal model SSL-GCN is hosted on an online server accessible through: https://app.cbbio.online/ssl-gcn/home.