PSRTTCA

PSRTTCA predicts and characterizes tumor T cell antigens (TTCAs) using a propensity score representation learning algorithm and a meta-predictor framework to improve machine-learning classifier performance.

Key Features:

• Possibility Score Representation Learning Algorithm: Generates multiple sets of propensity scores for amino acids, dipeptides, and g-gap dipeptides to determine their potential as TTCAs.
• Meta-predictor Framework Integration: Combines selected optimal sets of variant propensity scores into a framework that improves predictive accuracy compared to conventional methods.

Scientific Applications:

• Tumor T Cell Antigens Identification: Provides precise predictions for TTCAs, aiding understanding of their functional mechanisms and physicochemical properties.

Methodology:

Generates multiple sets of propensity scores for amino acids, dipeptides, and g-gap dipeptides via a propensity score representation learning algorithm and integrates selected optimal variant propensity-score sets into a meta-predictor framework to enhance machine-learning classifier performance for TTCAs.