Biology subjects

Morozova, I.

Publications and source records attributed to Morozova, I..

7 recordsLinked to original sources

bioRxiv2026

A Unified 3D Generative Model for Synthesizable Structure-Based Drug Design

Traditional screening-based drug discovery is inherently limited by the astronomical scale of the chemical space. Generative modelling offers a compelling alternative to the classical search paradigm and enables rational, bottom-up design of novel and target-specific small molecules. However, its impact has been hampered by challenges in synthetic accessibility of the designed compounds and lack of large-scale experimental validation. Here, we introduce LDDM (Large Drug Discovery Model), a generative framework that supports a range of drug discovery tasks, including constrained and unconstrained docking, fragment linking and growing, and de novo design. We further introduce a programmable design algorithm that enables accurate design of synthetically accessible compounds satisfying various fine-grained objectives. We experimentally validated the designed or optimised ligands for five therapeutically relevant protein targets. In all cases, LDDM achieved high success rates, allowing us to identify molecules with confirmed binding affinity while synthesizing only a small number of generated compounds. The best designs were structurally characterised through NMR spectroscopy and X-ray crystallography, demonstrating high prediction accuracy. Overall, LDDM provides a scalable and flexible platform for the rapid and tailored design of small molecules and non-natural peptides for therapeutic applications.

bioinformatics↗

bioRxiv2025

FragmentScope - exploring the fragment space with learned surface representations

Exploring fragment chemical space for ligand design remains a major challenge in early stage drug discovery. This task is particularly challenging due to the small size, low specificity, and weak binding affinities of low molecular weight (MW) fragments. We present FragmentScope, a computational pipeline that uses learned protein surface fingerprints to guide fragment placement and small molecule generation. By using a contrastive learning model trained on protein-ligand interactions, we built a database of surface-fragment pairs, which enables fast and accurate fragment placement in the target protein pocket. We demonstrate its effectiveness on benchmark datasets, achieving robust placement accuracy. FragmentScope also enables the design of small molecules based on the predicted fragments ensuring synthetic accessibility. We experimentally validated Fragmentscope across 5 different targets with binding assays and structural characterization. Our approach shows high success rates in fragment discovery and yielded promising leads for designed ligands. FragmentScope offers a scalable, structure-guided approach for narrowing chemical space and identifying prospective scaffolds, accelerating the early stages of small molecule design.

bioinformatics↗

bioRxiv2023

Exploring "dark matter" protein folds using deep learning

De novo protein design aims to explore uncharted sequence-and structure areas to generate novel proteins that have not been sampled by evolution. One of the main challenges in de novo design involves crafting "designable" structural templates that can guide the sequence search towards adopting the target structures. Here, we present an approach to learn patterns of protein structure based on a convolutional variational autoencoder, dubbed Genesis. We coupled Genesis with trRosetta to design sequences for a set of protein folds and found that Genesis is capable of reconstructing native-like distance-and angle distributions for five native folds and three novel, so-called "dark-matter" folds as a demonstration of generalizability. We used a high-throughput assay to characterize protease resistance of the designs, obtaining encouraging success rates for folded proteins and further biochemically characterized folded designs. The Genesis framework enables the exploration of the protein sequence and fold space within minutes and is not bound to specific protein topologies. Our approach addresses the backbone designability problem, showing that structural patterns in proteins can be efficiently learned by small neural networks and could ultimately contribute to the de novo design of proteins with new functions.

bioinformatics↗

bioRxiv2020

Leprosy in wild chimpanzees

Humans are considered the main host for Mycobacterium leprae, the aetiologic agent of leprosy, but spill-over to other mammals such as nine-banded armadillos and red squirrels occurs. Although naturally acquired leprosy has also been described in captive nonhuman primates, the exact origins of infection remain unclear. Here, we report on leprosy-like lesions in two wild populations of western chimpanzees (Pan troglodytes verus) in the Cantanhez National Park, Guinea-Bissau, and the Tai National Park, Cote dIvoire, West Africa. Longitudinal monitoring of both populations revealed the progression of disease symptoms compatible with advanced leprosy. Screening of faecal and necropsy samples confirmed the presence of M. leprae as the causative agent at each site and phylogenomic comparisons with other strains from humans and other animals show that the chimpanzee strains belong to different and rare genotypes (4N/O and 2F). The independent evolutionary origin of M. leprae in two geographically distant populations of wild chimpanzees, with no prolonged direct contact with humans, suggests multiple introductions of M. leprae from an unknown animal or environmental source.

evolutionary biology↗

bioRxiv2020

Sofosbuvir Terminated RNA is More Resistant to SARS-CoV-2 Proofreader than RNA Terminated by Remdesivir

SARS-CoV-2 is responsible for COVID-19, resulting in the largest pandemic in over a hundred years. After examining the molecular structures and activities of hepatitis C viral inhibitors and comparing hepatitis C virus and coronavirus replication, we previously postulated that the FDA-approved hepatitis C drug EPCLUSA (Sofosbuvir/Velpatasvir) might inhibit SARS-CoV-2.1 We subsequently demonstrated that Sofosbuvir triphosphate is incorporated by the relatively low fidelity SARS-CoV and SARS-CoV-2 RNA-dependent RNA polymerases (RdRps), serving as an immediate polymerase reaction terminator, but not by a host-like high fidelity DNA polymerase.2,3 Other investigators have since demonstrated the ability of Sofosbuvir to inhibit SARS-CoV-2 replication in lung and brain cells;4,5 additionally, COVID-19 clinical trials with EPCLUSA6 and with Sofosbuvir plus Daclatasvir7 have been initiated in several countries. SARS-CoV-2 has an exonuclease-based proofreader to maintain the viral genome integrity.8 Any effective antiviral targeting the SARS-CoV-2 RdRp must display a certain level of resistance to this proofreading activity. We report here that Sofosbuvir terminated RNA resists removal by the exonuclease to a substantially higher extent than RNA terminated by Remdesivir, another drug being used as a COVID-19 therapeutic. These results offer a molecular basis supporting the current use of Sofosbuvir in combination with other drugs in COVID-19 clinical trials.

pharmacology and toxicology↗

bioRxiv2020

Nucleotide Analogues as Inhibitors of SARS-CoV Polymerase

SARS-CoV-2, a member of the coronavirus family, has caused a global public health emergency.1 Based on our analysis of hepatitis C virus and coronavirus replication, and the molecular structures and activities of viral inhibitors, we previously reasoned that the FDA-approved heptatitis C drug EPCLUSA (Sofosbuvir/Velpatasvir) should inhibit coronaviruses, including SARS-CoV-2.2 Here, using model polymerase extension experiments, we demonstrate that the activated triphosphate form of Sofosbuvir is incorporated by low-fidelity polymerases and SARS-CoV RNA-dependent RNA polymerase (RdRp), and blocks further incorporation by these polymerases; the activated triphosphate form of Sofosbuvir is not incorporated by a host-like high-fidelity DNA polymerase. Using the same molecular insight, we selected two other anti-viral agents, Alovudine and AZT (an FDA approved HIV/AIDS drug) for evaluation as inhibitors of SARS-CoV RdRp. We demonstrate the ability of two HIV reverse transcriptase inhibitors, 3-fluoro-3-deoxythymidine triphosphate and 3-azido-3-deoxythymidine triphosphate (the active triphosphate forms of Alovudine and AZT), to be incorporated by SARS-CoV RdRp where they also terminate further polymerase extension. Given the 98% amino acid similarity of the SARS-CoV and SARS-CoV-2 RdRps, we expect these nucleotide analogues would also inhibit the SARS-CoV-2 polymerase. These results offer guidance to further modify these nucleotide analogues to generate more potent broad-spectrum anti-coronavirus agents.

pharmacology and toxicology↗

bioRxiv2019

Recent changes in candidate genes for domestication in humans in Europe: focusing on language

Human evolution resulted from changes in our biology, behavior, and culture. One source of these changes has been hypothesized to be our self-domestication (that is, the development in humans of features commonly found in domesticated strains of mammals, seemingly as a result of selection for reduced aggression). Signals of domestication, notably brain size reduction, have increased in recent times. In this paper we compare whole-genome data between Late Neolithic/Bronze Age individuals and modern Europeans and show that genes associated with mammal domestication and with neural crest development and function are significantly differently enriched in nonsynonymous single nucleotide polymorphisms between these two groups. We hypothesize how these changes might account for the increased features of self-domestication in modern humans and ultimately, for subtle recent changes in human cognition and behavior, including language.

evolutionary biology↗