We present a focused selection of 443 drug-like small molecules with machine learning-predicted antioxidant activity, refined through structure-based filtering and substructure exclusion. This Screening Set prioritizes molecules with known antioxidant pharmacophores (phenols, catechols, conjugated systems), while excluding PAINS and problematic scaffolds. It is designed to support early-stage drug discovery targeting oxidative stress-related diseases, including cancer, neurodegeneration and cardiovascular disorders.
The compound selection can be customized to meet your specific requirements. Cherry-picking is available.
Please, contact us at orders@lifechemicals.com for any additional information and price quotations.
Scientific Background
Reactive oxygen species (ROS) are highly reactive molecules that play a central role in cellular damage, aging and the pathogenesis of numerous other conditions, including cancer, neurodegenerative diseases and cardiovascular disorders.
Antioxidants, which neutralize ROS and reduce oxidative stress, hold significant promise as therapeutic agents [1-3]. Identifying new, structurally diverse antioxidants is critical for expanding the toolbox of bioactive molecules that can be optimized for bioavailability, potency and safety.
Computational methods have become increasingly vital in antioxidant research, as the high-throughput screening of large chemical libraries does enable identifying an incomparably greater number of potential candidates. By integrating cheminformatics filters and substructural analysis, we can prioritize molecules with a high likelihood of antioxidant activity while avoiding problematic or toxic motifs [4]. This approach enhances the efficiency of early-stage discovery and reduces downstream costs associated with synthesis and biological testing.
Compound selection
The compound selection began with machine learning (ML) predictions based on bioactivity data from validated assays (ABTS and DPPH) in ChEMBL (Fig. 1). A high-confidence threshold was used to minimize false positives, followed by an additional two-step refinement. Structure-based filtering, using RDKit’s PAINS and toxicophore alerts, removes undesirable scaffolds that are likely to yield false-positive biological signals [5]. Then, we flagged molecules with known antioxidant pharmacophores, such as phenol, catechol and conjugated diene structure motifs commonly associated with radical scavenging activity. The resulting compounds were ranked based on a custom scoring system, considering both antioxidant-relevant features and molecular weight. Thus , we have obtained a set of 443 smallmolecule compounds that are very well predicted to exhibit potent antioxidant activity.
Notably, the Screening Set features clusters of structurally similar compounds, as well as a broad distribution of chemically distinct molecules, as shown through dimensionality reduction analysis (Fig. 2). Accordingly, it provides a remarkable platform for both focused screening and the exploration of new chemical space in redox-related research.
Representative screening compounds from the Screening Library
Application prospects of our Antioxidant Screening Library:
- Identification of new antioxidant drug leads
- ROS-modulation screening projects
- Mechanism-of-action studies in oxidative stress
- Lead optimization and SAR analysis
- Functional studies in aging, inflammation, and neurodegeneration
Figure 1. (A) Pipeline of compound library selection. (B) Confusion matrix for selected cut-off value. (C) Structures of known antioxidant pharmacophores
Figure. 2. Representation of chemical space of selected antioxidants using the ChemPlot library and UMAP dimensionality reduction method.
Reference:
- Lobo, V et al. “Free radicals, antioxidants and functional foods: Impact on human health.” Pharmacognosy reviews vol. 4,8 (2010): 118-26. doi:10.4103/0973-7847.70902.
- Halliwell, Barry, and John M. C. Gutteridge, Free Radicals in Biology and Medicine, 5th edn (Oxford, 2015; online edn, Oxford Academic, 22 Oct. 2015), doi:10.1093/acprof:oso/9780198717478.001.0001.
- Alizadeh, Seyedeh Roya, and Mohammad Ali Ebrahimzadeh. “Quercetin derivatives: Drug design, development, and biological activities, a review.” European journal of medicinal chemistry vol. 229 (2022): 114068. doi:10.1016/j.ejmech.2021.114068.
- Lagorce, David et al. “Computational analysis of calculated physicochemical and ADMET properties of protein-protein interaction inhibitors.” Scientific reports vol. 7 46277. 11 Apr. 2017, doi:10.1038/srep46277.
- Baell, Jonathan B, and Georgina A Holloway. “New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays.” Journal of medicinal chemistry vol. 53,7 (2010): 2719-40. doi:10.1021/jm901137j.