ChloroP

ChloroP predicts chloroplast transit peptides and their cleavage sites in plant proteins to identify chloroplast-targeting signals.

Key Features:

• Neural Network-Based Prediction: Employs a neural network trained on a homology-reduced, curated dataset to distinguish transit peptides from non-transit peptides.
• High Accuracy: Achieves an 88% correct classification rate on its training set and exceeds the performance reported for PSORT.
• Cleavage Site Prediction: Predicts cleavage sites using a scoring matrix generated by an automatic motif-finding algorithm, approximating known cleavage sites with approximately 60% accuracy within ±2 residues based on SWISS-PROT data.

Scientific Applications:

• Genome-Wide Analysis: Enables genome-wide identification of putative transit peptides across large sequence datasets, demonstrated on 715 Arabidopsis thaliana sequences from SWISS-PROT.
• Research and Development: Supports functional genomics, protein engineering, and synthetic biology studies that require prediction of chloroplast-targeting signals.

Methodology:

Uses a neural network trained on homology-reduced, curated datasets to identify transit peptides; cleavage-site prediction is based on a scoring matrix produced by an automatic motif-finding algorithm.