PORTRAIT

PORTRAIT predicts non-coding RNAs in transcriptome sequences, enabling identification of ncRNAs from low-quality or incomplete transcripts and in datasets lacking protein homologs.

Key Features:

• Error-Resilient Translation: Uses an error-tolerant translation step to generate putative proteins from transcript sequences despite sequencing errors and incomplete open reading frames.
• Support Vector Machine (SVM) Evaluation: Evaluates coding potential with SVM models, applying a protein-dependent model when a putative protein is identified and a protein-independent model when none is found.
• Ab Initio Feature Extraction: Extracts exclusively ab initio features and does not rely on homology information.
• Integration and Efficiency: Designed for integration into bioinformatics pipelines and optimized for low computational cost for large-scale ncRNA detection.

Scientific Applications:

• Transcriptome ncRNA discovery: Identification of ncRNAs in transcriptomes from neglected or poorly characterized species where genomic resources and protein homology are limited.
• Fungal pathogen analysis: Application to the transcriptome of Paracoccidoides brasiliensis and five related fungi to predict ncRNAs in complex pathogenic fungal datasets.

Methodology:

Translate transcript sequences with error-tolerant software, extract ab initio features, and assess coding potential using dual SVM models (protein-dependent if a putative protein is present; protein-independent if not).