Bioscientifica (United Kingdom)

The properties of a new drug are set at the point of design, including any potential design flaws. Many small molecules that make it to human trials fail to progress beyond Phase 1 due to these flaws, which manifest as unacceptable side effects (either alone or in combination), meaning the drug (and therefore the pathway it is modulating) can never be adequately evaluated. This presentation will highlight how, through the thoughtful generation of a precise target candidate profile (TCP) and the subsequent use of GenAI, clinical candidates can now be precision designed to more exacting standards. Such molecules have the potential to have improved therapeutic indices, even against protein targets that are usually associated with mechanism-based adverse events. Citation Format: David J. Hallett. Overcoming traditional design limitations with AI-based discovery. [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Optimizing Therapeutic Efficacy and Tolerability through Cancer Chemistry; 2024 Dec 9-11; Toronto, Ontario, Canada. Philadelphia (PA): AACR; Mol Cancer Ther 2024;23(12_Suppl):Abstract nr IA004

Sire is a Python/C++ library that is used both to prototype new algorithms and as an interoperability engine for exchanging information between molecular simulation programs. It provides a collection of file parsers and information converters that together make it easier to combine and leverage the functionality of many other programs and libraries. This empowers researchers to use sire to write a single script that can, for example, load a molecule from a PDBx/mmCIF file via Gemmi, perform SMARTS searches via RDKit, parameterize molecules using BioSimSpace, run GPU-accelerated molecular dynamics via OpenMM, and then display the resulting dynamics trajectory in a NGLView Jupyter notebook 3D molecular viewer. This functionality is built on by BioSimSpace, which uses sire's molecular information engine to interconvert with programs such as GROMACS, NAMD, Amber, and AmberTools for automated molecular parameterization and the running of molecular dynamics, metadynamics, and alchemical free energy workflows. Sire comes complete with a powerful molecular information search engine, plus trajectory loading and editing, analysis, and energy evaluation engines. This, when combined with an in-built computer algebra system, gives substantial flexibility to researchers to load, search for, edit, and combine molecular information from multiple sources and use that to drive novel algorithms by combining functionality from other programs. Sire is open source (GPL3) and is available via conda and at a free Jupyter notebook server at https://try.openbiosim.org. Sire is supported by the not-for-profit OpenBioSim community interest company.

C–H borylation is a high-value transformationin the synthesisof lead candidates for the pharmaceutical industry because a widearray of downstream coupling reactions is available. However, predictingits regioselectivity, especially in drug-like molecules that may containmultiple heterocycles, is not a trivial task. Using a data set ofborylation reactions from Reaxys, we explored how a language modeloriginally trained on USPTO_500_MT, a broad-scope set of patent data,can be used to predict the C–H borylation reaction productin different modes: product generation and site reactivity classification.Our fine-tuned T5Chem multitask language model can generate the correctproduct in 79% of cases. It can also classify the reactive aromaticC–H bonds with 95% accuracy and 88% positive predictive value,exceeding purpose-developed graph-based neural networks.

DNA-encoded libraries (DELs) have become a leading technology for hit identification in drug discovery projects as large, diverse libraries can be generated. DELs are commonly synthesised via split-and-pool methodology; thus, chemical transformations utilised must be highly efficient, proceeding with high conversions. Reactions performed in DEL synthesis also require a broad substrate scope to produce diverse, drug-like libraries. Many pharmaceutical compounds incorporate multiple C-N bonds, over a quarter of which are synthesised via reductive aminations. However, few on-DNA reductive amination procedures have been developed. Herein is reported the application of the micelle-forming surfactant, TPGS-750-M, to the on-DNA reductive amination of DNA-conjugated amines, yielding highly efficient conversions with a broad range of aldehydes, including medicinally relevant heterocyclic and aliphatic substrates. The procedure is compatible with DNA amplification and sequencing, demonstrating its applicability to DEL synthesis.

The European Lead Factory (ELF) is a consortium of universities and small and medium-sized enterprises (SMEs) dedicated to drug discovery, and the pharmaceutical industry. This unprecedented consortium provides high-throughput screening, triage, and hit validation, including to non-consortium members. The ELF library was created through a novel compound-sharing model between nine pharmaceutical companies and expanded through library synthesis by chemistry-specialized SMEs. The library has been screened against ∼270 different targets and 15 phenotypic assays, and hits have been developed to form the basis of patents and spin-off companies. Here, we review the outcome of screening campaigns of the ELF, including the performance and physicochemical properties of the library, identification of possible frequent hitter compounds, and the effectiveness of the compound-sharing model.

FtsZ is an essential bacterial protein abundantly studied as a novel and promising target for antimicrobials. FtsZ is highly conserved among bacteria and mycobacteria, and it is crucial for the correct outcome of the cell division process, as it is responsible for the division of the parent bacterial cell into two daughter cells. In recent years, the benzodioxane–benzamide class has emerged as very promising and capable of targeting both Gram-positive and Gram-negative FtsZs. In this study, we explored the effect of including a substituent on the ethylenic linker between the two main moieties on the antimicrobial activity and pharmacokinetic properties. This substitution, in turn, led to the generation of a second stereogenic center, with both erythro and threo isomers isolated, characterized, and evaluated. With this work, we discovered how the hydroxy group slightly affects the antimicrobial activity, while being an important anchor for the exploitation and development of prodrugs, probes, and further derivatives.

Abstract The analysis of extensive electronic health records (EHR) datasets often calls for automated solutions, with machine learning (ML) techniques, including deep learning (DL), taking a lead role. One common task involves categorizing EHR data into predefined groups. However, the vulnerability of EHRs to noise and errors stemming from data collection processes, as well as potential human labeling errors, poses a significant risk. This risk is particularly prominent during the training of DL models, where the possibility of overfitting to noisy labels can have serious repercussions in healthcare. Despite the well-documented existence of label noise in EHR data, few studies have tackled this challenge within the EHR domain. Our work addresses this gap by adapting computer vision (CV) algorithms to mitigate the impact of label noise in DL models trained on EHR data. Notably, it remains uncertain whether CV methods, when applied to the EHR domain, will prove effective, given the substantial divergence between the two domains. We present empirical evidence demonstrating that these methods, whether used individually or in combination, can substantially enhance model performance when applied to EHR data, especially in the presence of noisy/incorrect labels. We validate our methods and underscore their practical utility in real-world EHR data, specifically in the context of COVID-19 diagnosis. Our study highlights the effectiveness of CV methods in the EHR domain, making a valuable contribution to the advancement of healthcare analytics and research.

Antibodies with similar amino acid sequences, especially across their complementarity-determining regions, often share properties. Finding that an antibody of interest has a similar sequence to naturally expressed antibodies in healthy or diseased repertoires is a powerful approach for the prediction of antibody properties, such as immunogenicity or antigen specificity. However, as the number of available antibody sequences is now in the billions and continuing to grow, repertoire mining for similar sequences has become increasingly computationally expensive. Existing approaches are limited by either being low-throughput, non-exhaustive, not antibody specific, or only searching against entire chain sequences. Therefore, there is a need for a specialized tool, optimized for a rapid and exhaustive search of any antibody region against all known antibodies, to better utilize the full breadth of available repertoire sequences. We introduce Known Antibody Search (KA-Search), a tool that allows for the rapid search of billions of antibody variable domains by amino acid sequence identity across either the variable domain, the complementarity-determining regions, or a user defined antibody region. We show KA-Search in operation on the sim 2.4 billion antibody sequences available in the OAS database. KA-Search can be used to find the most similar sequences from OAS within 30 minutes and a representative subset of 10 million sequences in less than 9 seconds. We give examples of how KA-Search can be used to obtain new insights about an antibody of interest. KA-Search is freely available at https://github.com/oxpig/kasearch.