ODNA

ODNA classifies organellar (mitochondrial and chloroplast) DNA sequences within whole genome assemblies to distinguish them from nuclear DNA.

Key Features:

• Machine learning classification: Uses a machine-learning model trained on 829,769 DNA sequences from 405 genome assemblies to identify organellar sequences within genomic data.
• Predictive performance: Reports Matthew's correlation coefficients (MCC) of 0.61 for mitochondrial sequences and 0.73 for chloroplast sequences on independent validation datasets.
• Genome annotation workflow: Applies a predefined genome annotation workflow to systematically process input assemblies prior to classification.

Scientific Applications:

• Assembly organelle identification: Distinguishing organellar DNA from nuclear DNA within whole genome assemblies.
• Evolutionary biology: Analyzing organelle sequence variation and evolutionary dynamics.
• Plant and animal genetics: Identifying chloroplast and mitochondrial sequences relevant to genetic studies in plants and animals.
• Comparative genomics: Enabling comparative analyses that require separation of organellar and nuclear sequences.

Methodology:

Predefined genome annotation workflow processes input assemblies, and a machine-learning model trained on 829,769 DNA sequences from 405 genome assemblies performs the organellar sequence classification.