The DEL Wall

Machine learning on DNA-Encoded Library datasets — one of the most exciting frontiers in drug discovery.

What is a DEL?

A DNA-Encoded Library (DEL) is a collection of millions (or billions) of small molecules, each tagged with a unique DNA barcode. Instead of testing molecules one at a time, you can screen the entire library at once against a protein target — then use DNA sequencing to figure out which molecules bound.

The result? Massive datasets with hundreds of millions of molecules, each with some signal about whether it binds. The ML challenge is separating real binders from noise, especially when structurally similar molecules can have very different binding behavior.

Before starting these routes, ask your chatbot:

• "What is a DNA-Encoded Library and how is it used in drug discovery?"
• "How does affinity selection work in DEL screening?"
• "What does 'enrichment' mean in the context of DEL data?"

What You'll Learn

• How DEL screens work and why they generate massive datasets
• Three approaches to molecular classification: frozen embeddings → fine-tuning → contrastive learning
• Working with real kinase drug discovery data
• Why hard negatives matter for learning meaningful features

Why This Wall Matters

DEL screens let you test hundreds of millions of molecules at once. The ML challenge is figuring out which ones actually bind. These routes teach you to think critically about how you frame learning problems — classification vs. contrastive, easy vs. hard negatives.

Warning: this wall is steep. The routes here assume you're comfortable with ML fundamentals from earlier walls.