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

Area of Science:

• Protein engineering and antibody development.
• High-throughput screening and deep sequencing analysis.
• Computational biology and bioinformatics.

Background:

• Deep sequencing of protein libraries after display sorting is crucial for identifying improved variants.
• Conventional methods using frequency or enrichment ratios often miss optimal variants and underutilize data.
• There is a need for more robust methods to analyze deep sequencing data for protein engineering.

Approach:

• Developed Position-Specific Enrichment Ratio Matrix (PSERM) scoring using entire deep sequencing datasets.
• PSERM scores sum site-specific enrichment ratios at each mutated position.
• Applied PSERM to analyze antibody variants, including a clinical-stage antibody (emibetuzumab).

Key Points:

• PSERM scores demonstrate higher reproducibility compared to traditional methods.
• PSERM scores show stronger correlation with experimentally measured antibody properties (affinity, non-specific binding).
• The method effectively utilizes comprehensive deep sequencing data for variant selection.

Conclusions:

• PSERM scoring is a more effective approach for identifying optimal protein variants from deep sequencing data.
• This method enhances the analysis of antibody engineering campaigns and related protein engineering efforts.
• The PSERM method and associated code are publicly available for broad application.