Nanobody Polyreactivity Prediction Dataset

Dataset Overview

This dataset helps predict whether nanobodies will show polyreactivity - the tendency to bind to multiple unrelated antigens. Polyreactivity is usually an unwanted feature in therapeutic applications, as it can lead to side effects and reduced effectiveness.

Accurately predicting nanobody polyreactivity is important for screening high-quality therapeutic candidates and understanding the molecular basis of antibody specificity.

Data Collection

The dataset is based on nanobody polyreactivity data measured in laboratory experiments and we collect it from public literature. Nanobodies are classified as polyreactive or non-polyreactive based on their binding to these different antigens.

Dataset Structure

The dataset is split into training, validation, and test sets.

File Format

CSV files contain these columns:

• seq: Nanobody amino acid sequence
• label: Binary label indicating polyreactivity (1 for high polyreactivity, 0 for low polyreactivity)

Uses and Limitations

Uses

• Develop models to predict nanobody polyreactivity
• Screen for highly specific nanobody candidates
• Understand sequence features and molecular basis of polyreactivity

Limitations

• Experimental methods for measuring polyreactivity may vary
• Polyreactivity exists on a spectrum rather than as a strict binary property
• Different experimental conditions may affect polyreactivity

Evaluation Metrics

Model performance is evaluated using:

• Accuracy
• F1 Score
• Precision
• Recall
• AUROC (Area Under the Receiver Operating Characteristic curve)
• AUPRC (Area Under the Precision-Recall Curve)