SeqScreen

SeqScreen assigns taxonomic classifications, functional annotations, and biological-process labels to short DNA sequences using ensemble machine learning to enable functionally informed pathogen characterization.

Key Features:

• Ensemble machine learning model: Uses an ensemble learning approach to label protein-coding sequences.
• Functions of Sequences of Concern (FunSoCs): Applies a specialized, curated set of FunSoCs tailored to microbial pathogenesis for functional labeling.
• Taxonomic classification: Assigns taxonomic classifications to short DNA sequences.
• Functional annotation: Produces functional annotations and biological-process labels for short DNA sequences.
• Integration of labels: Integrates taxonomic and functional labels for combined analysis beyond sequence identification.
• Performance: Reports high sensitivity, recall, and precision in its sequence annotations.

Scientific Applications:

• Pathogen detection and characterization: Identification and characterization of known and emerging pathogens from sequence data.
• Clinical and environmental surveillance: Analysis of clinical and environmental samples to detect pathogenic sequences.
• Microbial ecology and public health: Informing microbial ecology studies and public health surveillance with functionally informed annotations.

Methodology:

SeqScreen applies an ensemble machine learning model to label protein-coding sequences using curated FunSoCs and integrates taxonomic and functional labels to assign taxonomic classifications, functional annotations, and biological-process labels to short DNA sequences.