Modern computational approaches and machine learning techniques accelerate the invention of new drugs. Generative models can discover novel molecular structures within hours, while conventional drug discovery pipelines require months of work. In this article, we propose a new generative architecture, entangled conditional adversarial autoencoder, that generates molecular structures based on various properties, such as activity against a specific protein, solubility, or ease of synthesis. We apply the proposed model to generate a novel inhibitor of Janus kinase 3, implicated in rheumatoid arthritis, psoriasis, and vitiligo. The discovered molecule was tested in vitro and showed good activity and selectivity.
Plasma membrane is one of the major targets for cationic antiseptics (CA). The study was aimed to assess molecular effects of CAs of different chemical classes on cardiolipin-containing regions of bacterial plasma membranes. The study was carried out using coarse-grained molecular modeling. Interaction of CAs, such as miramistin, chlorhexidine, picloxidine, and octenidine, with cardiolipin-containing bilayer was assessed based on the CA coarse-grained models. CAs reduced lipid lateral diffusion coefficients and increased the membrane area per lipid. All CAs, except miramistin, reduced the lipid fatty acid chain order parameters. Adding octenidine at a CA : lipid ratio of 1 : 4 resulted in cardiolipin clustering with subsequent pulling the neutral phosphatidylethanolamine molecules out of the model bilayer. It was found that CАs have the potential for sorption to lipid bilayer, causing clustering of negatively charged lipids. Antiseptic octenidine causes formation of cardiolipin microdomains. Abnormal lateral lipid distribution together with pulling out phosphatidylethanolamine molecules can result in increased lipid bilayer permeability. The most significant reduction of cardiolipin lateral diffusion coefficient by 2.8 ± 0.4 times was observed in the presence of CA chlorhexidine at an antiseptic : lipid ratio of 1 : 4.
SARS-CoV-2 causing coronavirus disease 2019 (COVID-19) is responsible for one of the most deleterious pandemics of our time. The interaction between the ACE2 receptors at the surface of human cells and the viral Spike (S) protein triggers the infection making the receptor-binding domain (RBD) of the SARS-CoV-2 S-protein a focal target for the neutralizing antibodies (Abs). Despite the recent progress in the development and deployment of vaccines, the emergence of novel variants of SARS-CoV-2 insensitive to Abs produced in response to the vaccine administration and/or monoclonal ones represents upcoming jeopardy. Here, we assessed the possible effects of single and multiple mutations in the RBD of SARS-CoV-2 S-protein on its binding energy to various antibodies and the human ACE2 receptor. The performed computational analysis indicates that while single amino acid replacements in RBD may only cause partial impairment of the Abs binding, moreover, limited to specific epitopes, some variants of SARS-CoV-2 (with as few as 8 mutations), which are already present in the population, may potentially result in a much broader antigenic escape. We also identified a number of point mutations, which, in contrast to the majority of replacements, reduce RBD affinity to various antibodies without affecting its binding to ACE2. Overall, the results provide guidelines for further experimental studies aiming at the identification of the high-risk RBD mutations allowing for an antigenic escape.