PRELYM (Prediction of Lysine Modification) is a computational tool that predicts which lysine residues and N-terminal amino groups in a protein will react with bioconjugation reagents, and in what order. By analyzing a protein's three-dimensional structure, PRELYM empowers researchers to design more effective protein modifications for therapeutic and biotechnology applications.

Protein modification is fundamental to modern biotechnology, from developing longer-lasting therapeutics to creating enzyme-polymer conjugates with enhanced stability. However, predicting where and how proteins will be modified has historically relied on costly trial-and-error experiments.

PRELYM transforms this process by using structure-based prediction. The program analyzes several key molecular properties:

• Surface Accessibility: How exposed each amino group is to potential modification
• Chemical Environment: The pKa, local charge, and surrounding chemical interactions
• Secondary Structure: Whether the residue is in a helix, strand, or coil
• Quaternary Structure: How protein oligomerization affects modification patterns

By integrating these parameters through a validated decision tree algorithm, PRELYM classifies each reactive site as fast-reacting, slow-reacting, or non-reacting. This automated analysis takes just seconds and has been validated across multiple proteins including lysozyme, chymotrypsin, glucose oxidase, avidin, and therapeutic proteins like interferon-α 2a.

Origins in ATRP Research

PRELYM emerged from extensive research in Atom Transfer Radical Polymerization (ATRP)—a technique for growing polymers directly from protein surfaces. Unlike traditional PEGylation, ATRP offers precise control over polymer grafting density and location, enabling the creation of protein-polymer conjugates with dramatically improved stability, activity retention, and therapeutic properties.

Through systematic characterization of how ATRP initiators react with protein surfaces, the research team developed predictive rules based on tertiary structure analysis. PRELYM automates this knowledge, eliminating the laborious manual calculations previously required.