ALLERDET

ALLERDET predicts the allergenic potential of proteins from genetically modified organisms (GMOs) and novel foods using machine-learning models and sequence-alignment-based feature extraction to assess food allergenicity.

Key Features:

• Artificial Intelligence Integration: Employs Restricted Boltzmann Machines and Decision Trees to model complex patterns in allergenicity data.
• Pairwise Sequence Alignment (FASTA): Uses the FASTA program to perform pairwise sequence alignment for feature extraction by comparing query proteins to known allergens.
• Performance Metrics: Reports sensitivity of 98.46%, specificity of 94.37%, and overall accuracy of 97.26% for predictive performance.

Scientific Applications:

• Food Safety Assessment: Predicts allergenic potential of proteins derived from GMOs or novel foods to inform safety evaluations.
• Research and Development: Screens new protein variants for potential allergenicity during product development.
• Regulatory Compliance: Supports evaluation of genetically modified crops and novel proteins for allergenicity in regulatory assessments.

Methodology:

Features are extracted via pairwise sequence alignment using the FASTA program and used as input to a hybrid model combining Restricted Boltzmann Machines and Decision Trees.