The role of human in the loop: lessons from D3R challenge 4

Computation-Directed Molecular Design, Synthesis, and Fungicidal Activity of Succinate Dehydrogenase Inhibitors

Succinate dehydrogenase inhibitors (SDHIs) are a class of fungicides targeting the pathogenic fungi mitochondrial SDH. Here, molecular docking, three-dimensional quantitative structure-activity relationship (3D-QSAR), and molecular dynamics (MD) simulations were used to guide SDHI innovation. Molecular docking was performed to explore the binding modes of SDH and its inhibitors. 3D-QSAR models were carried out on 33 compounds with activity against Rhizoctonia cerealis (R. cerealis); their structure-activity relationships were analyzed using comparative molecular field analysis and comparative molecular similarity indices analysis. MD simulations were used to assess the stability of the complexes under physiological conditions, and the results were consistent with molecular docking. Binding free energy was calculated through the molecular mechanics generalized born surface area method, and the binding free energy was decomposed. The results are consistent with the activity of bioassay and indicate that van der Waals and lipophilic interactions contribute the most in the molecular binding process. Afterward, we designed and synthesized 12 compounds under the guidance of the above-mentioned analyses, bioassay found that F9 was active against R. cerealis with the EC50 value of 9.43 μg/mL, and F4, F5, and F9 were active against Botrytis cinerea with an EC50 values of 5.80, 3.17, and 1.63 μg/mL, respectively. They all showed good activity between positive controls of pydiflumetofen and thifluzamide. Our study provides new considerations for effective SDHIs discovery.

A New Class of Uracil-DNA Glycosylase Inhibitors Active against Human and Vaccinia Virus Enzyme

Uracil-DNA glycosylases are enzymes that excise uracil bases appearing in DNA as a result of cytosine deamination or accidental dUMP incorporation from the dUTP pool. The activity of Family 1 uracil-DNA glycosylase (UNG) activity limits the efficiency of antimetabolite drugs and is essential for virulence in some bacterial and viral infections. Thus, UNG is regarded as a promising target for antitumor, antiviral, antibacterial, and antiprotozoal drugs. Most UNG inhibitors presently developed are based on the uracil base linked to various substituents, yet new pharmacophores are wanted to target a wide range of UNGs. We have conducted virtual screening of a 1,027,767-ligand library and biochemically screened the best hits for the inhibitory activity against human and vaccinia virus UNG enzymes. Although even the best inhibitors had IC₅₀ ≥ 100 μM, they were highly enriched in a common fragment, tetrahydro-2,4,6-trioxopyrimidinylidene (PyO3). In silico, PyO3 preferably docked into the enzyme's active site, and in kinetic experiments, the inhibition was better consistent with the competitive mechanism. The toxicity of two best inhibitors for human cells was independent of the presence of methotrexate, which is consistent with the hypothesis that dUMP in genomic DNA is less toxic for the cell than strand breaks arising from the massive removal of uracil. We conclude that PyO3 may be a novel pharmacophore with the potential for development into UNG-targeting agents.

Preliminary modelling as the first stage of targeted organic synthesis

The review aims to present a classification and applicability analysis of methods for preliminary molecular modelling for targeted organic, catalytic and biocatalytic synthesis. The following three main approaches are considered as a primary classification of the methods: modelling of the target – ligand coordination without structural information on both the target and the resulting complex; calculations based on experimentally obtained structural information about the target; and dynamic simulation of the target – ligand complex and the reaction mechanism with calculation of the free energy of the reaction. The review is meant for synthetic chemists to be used as a guide for building an algorithm for preliminary modelling and synthesis of structures with specified properties.

The bibliography includes 353 references.

A Review on Human-AI Interaction in Machine Learning and Insights for Medical Applications

Objective: To provide a human–Artificial Intelligence (AI) interaction review for Machine Learning (ML) applications to inform how to best combine both human domain expertise and computational power of ML methods. The review focuses on the medical field, as the medical ML application literature highlights a special necessity of medical experts collaborating with ML approaches. Methods: A scoping literature review is performed on Scopus and Google Scholar using the terms “human in the loop”, “human in the loop machine learning”, and “interactive machine learning”. Peer-reviewed papers published from 2015 to 2020 are included in our review. Results: We design four questions to investigate and describe human–AI interaction in ML applications. These questions are “Why should humans be in the loop?”, “Where does human–AI interaction occur in the ML processes?”, “Who are the humans in the loop?”, and “How do humans interact with ML in Human-In-the-Loop ML (HILML)?”. To answer the first question, we describe three main reasons regarding the importance of human involvement in ML applications. To address the second question, human–AI interaction is investigated in three main algorithmic stages: 1. data producing and pre-processing; 2. ML modelling; and 3. ML evaluation and refinement. The importance of the expertise level of the humans in human–AI interaction is described to answer the third question. The number of human interactions in HILML is grouped into three categories to address the fourth question. We conclude the paper by offering a discussion on open opportunities for future research in HILML.

Computational modeling and target synthesis of monomethoxy-substituted o-diphenylisoxazoles with unexpectedly high antimitotic microtubule destabilizing activity

The ability of monomethoxy-substituted o-diphenylisoxazoles 2a-d to interact with the colchicine site of tubulin was predicted using computational modeling, docking studies, and calculation of binding affinity. The respective molecules were synthesized in high yields by three steps reaction using easily available benzaldehydes, acetophenones, and arylnitromethanes as starting material. The calculated antitubulin effect was confirmed in vivo in a sea urchin embryo model. Compounds 2a and 2c showed high antimitotic microtubule destabilizing activity compared to that of CA4. Isoxazole 2a also exhibited significant cytotoxicity against human cancer cells in NCI60 screen. For the first time, isoxazole-linked CA4 derivatives 2a and 2c with only one methoxy substituent were identified as potent antimitotic microtubule destabilizing agents. These molecules could be considered as promising structures for further optimization.

Computational identification of disulfiram and neratinib as putative SARS-CoV-2 main protease inhibitors

Identification of disulfiram and neratinib as putative covalent inhibitors of SARS-CoV-2 virus main protease M^pro by a combination of ‘on-top docking’ procedure, expert evaluation of potential hits and molecular dynamics is reported herein. This finding shows the importance of further development of virtual screening add-ons.