Position-Specific Enrichment Ratio Matrix scores predict antibody variant properties from deep sequencing data

Area of Science:

• Protein engineering
• Computational biology
• Biophysics

Background:

• Deep sequencing of protein libraries after display sorting is crucial for identifying improved variants.
• Current methods using variant frequencies or enrichment ratios often fail to identify the best candidates and disregard significant data.

Purpose of the Study:

• To introduce a novel scoring method, Position-Specific Enrichment Ratio Matrix (PSERM), for analyzing deep sequencing data.
• To improve the identification of optimal protein variants by utilizing the entire deep sequencing dataset.

Main Methods:

• The PSERM scoring method aggregates site-specific enrichment ratios across all mutated positions for each variant.
• It utilizes complete deep sequencing datasets from pre- and post-selection stages.

Main Results:

• PSERM scores demonstrate higher reproducibility compared to traditional frequency and enrichment ratio methods.
• PSERM scores show stronger correlation with experimentally measured antibody properties, including affinity and non-specific binding, for emibetuzumab.
• The method proved effective for a clinical-stage antibody, suggesting broad applicability.

Conclusions:

• PSERM scoring offers a more robust and comprehensive approach to analyzing deep sequencing data in protein engineering.
• This method enhances the identification of superior protein variants for various applications.
• The PSERM method is expected to be widely adopted in diverse protein engineering campaigns.