Last week we launched an exciting initiative at PEGS Europe 2024 to help advance antibody discovery using AI and open source data. Our goal with this is to empower more builders by funding and creating high-quality open source NGS datasets for the community.

Below is a closer look at what we shared at PEGS Europe and how you can get involved.

Submission link: https://cradlebio.typeform.com/targets

The Opportunity: Better Drugs

In recent experiments we’ve seen great potential in using NGS data combined with ML for antibody optimization projects. By using an NGS (next-generation sequencing) panning dataset focused on SARS-CoV-2, our models were able to learn and generate sequences with up to 30% amino acid variation—all while maintaining strong binding affinity to target proteins. This result marks a significant step forward: it shows that AI models trained on NGS data can go beyond the initial sequence, outperforming those trained only on validated binders. The takeaway? NGS data could be a game-changer in accelerating AI-driven drug discovery, pushing forward new therapeutic options faster and more effectively.

The Disease: Lack of Data

One major obstacle to doing this at scale is the lack of comprehensive datasets. While NGS data holds tremendous potential, there’s a severe shortage of public datasets that include comprehensive NGS data for every round. This shortage creates a very real bottleneck, limiting both the development of new models as well as the field of drug discovery in general.

We're solving this bottleneck by building a high-quality and open data resource. By inviting the scientific community to contribute to this effort we hope to also maximize its impact on the industry.

The Cure: Your Help

Now we’re asking you, researchers and scientists, to join us in accelerating antibody discovery and, ultimately, the development of better drugs. You can help by submitting potential antibody targets that meet our checklist:

• protein or peptide

• is soluble

• is stable

• is safe to handle

• is easily manufacturable

• mono- or multi-specific applications

• ideally underserved area*

By sharing target ideas that fit these criteria, you’ll contribute to an open pool for NGS data that could accelerate AI-based discovery further. We’ll take care of the experiments and publish the resulting data for anyone to use and build upon.

*it would be amazing to build datasets in areas like orphan or poverty-related diseases

How to Contribute: 

Simply go to https://cradlebio.typeform.com/targets and submit your target ideas. Every submission moves us closer to a future where faster, more effective antibody discoveries are within reach.