Background:
We aim to unlock the therapeutic potential of 'undruggable' targets. Macrocyclic peptides are uniquely positioned to engage intracellular targets that are inaccessible to most biologics, and they also provide the potential to address targets without a ligand-able pocket, which small molecule therapeutics require. Macrocyclic peptides represent a promising therapeutic modality due to their potential for high target specificity, increased stability and ability to modulate challenging biological targets. However, achieving oral bioavailability and/or sufficient cell permeability remains a key challenge in the development and deployment of macrocyclic peptides for clinical use.
Critical to the next wave of peptide innovation is a comprehensive understanding of the property space that governs oral absorption and cell permeability for macrocyclic peptides. Data-driven approaches, AI/ML augmentation, and high-throughput experimental platforms are central to our strategy to accelerate discovery and enhance our ability to design and deliver breakthrough peptide therapies.
To address this, AstraZeneca seeks to harness the power of Open Innovation by inviting collaborators from academia, biotech, and industry to share existing data sets or compound collections for screening through our in house cascade to determine physicochemical, permeability and stability properties of macrocyclic peptides. Leveraging AstraZeneca's expertise in molecular data analytics and AI, the collaborative team will use these data sets/compounds to develop multiparameter AI/ML workflows to identify the most valuable predictive screening methods and enable robust prediction of the properties required for oral and cell-permeable macrocyclic peptides.
The Challenge:
AstraZeneca invites partners across the global scientific community to contribute datasets or physical/virtual libraries of macrocyclic peptides that can help map and model the property space required for oral and cell-permeable macrocyclic peptides. We particularly welcome:
- Annotated in vitro/vivo datasets describing permeability, oral bioavailability, enzymatic stability or other relevant properties of peptide macrocycles—including both proprietary and published collections.
- Libraries of macrocyclic peptides which are suitable for screening, ideally with reference activity which would enable AI-led property mapping and predictive modelling. Physical compound sets are preferred, but virtual libraries may be accepted for use once training datasets have been established.
- Novel or proprietary methods for predicting, quantifying, or modelling the physicochemical space that governs permeability and oral absorption in macrocyclic peptides
Proposed collaborations or data contributions should either provide experimental proof-of-concept, published evidence, or actionable approaches for generating datasets or screening tools that address these key parameters. Solutions that enable parallel evaluation of multiple macrocyclic peptides, or which deliver mechanistic or predictive insights into the determinants of oral and cell permeability, are particularly encouraged.
The Solution:
We are looking for collaborative solutions. To be considered for partnership or funding, proposals should demonstrate:
- Existing availability of experimental datasets or chemical macrocycle libraries, applicable to oral and cell-permeable macrocyclic peptides profiling (i.e. the research plan cannot be hypothetical in nature).
- A clear description of technology readiness or the immediate feasibility for transfer and use by AstraZeneca or its partners.
- Evidence of relevance, such as proof-of-concept experiments, published data, prior deployment in peptide profiling or optimisation, or demonstrated utility in property mapping/predictive modelling.
- For data sets, the potential for standardisation (FAIR data) from documented screening methods that generate reliable and reproduceable data to allow scalable integration and validation within AstraZeneca developed AI/ML models.
- Description of supporting expertise, infrastructure, or computational resources enabling rapid onboarding or further development.
Out of scope:
Proposals that are hypothetical, lack experimental feasibility, or involve low-throughput experimental or predictive modelling approaches that are not scalable for parallel evaluation of large compound sets will be given a lower priority.
Benefits of collaborators:
- Co-development opportunity with AstraZeneca
- Access to comprehensive screening data: contributors will receive complete physicochemical, permeability, and stability profiles for all compounds screened through AstraZeneca's validated cascade
- AI/ML model insights: collaborators gain access to predictive models and property-activity relationships derived from their contributed data, enhancing their own research capabilities
- Co-publication opportunities: joint publication rights for significant findings and model developments arising from collaborative datasets
- Expanded dataset value: data becomes part of a larger, more statistically robust dataset, increasing its scientific impact and citation potential.
