New method estimates DNA-encoded library size for machine learning models

Researchers have developed a new method to estimate the required size of a second-generation DNA-encoded library (DEL) for training high-resolution binding models. This approach uses an "energy pixels" concept, treating binding as a sum of small interactions between a target protein and a ligand. By analyzing the spread of affinities from an initial DEL, the method determines the number of compounds needed for dense sampling and identifies tight binders, thereby estimating the complexity required for effective machine learning guidance in medicinal chemistry. AI

IMPACT This research could improve the efficiency of drug discovery by optimizing the training data for machine learning models used in medicinal chemistry.

RANK_REASON The cluster contains a research paper detailing a new methodology for training machine learning models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Hugging Face Daily Papers →

• active learning
• Delphinus
• ligand
• machine learning model
• medicinal chemistry
• PIXELS

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →