krocus

Krocus predicts multi-locus sequence types (MLST) directly from uncorrected long-read sequencing data to enable rapid bacterial strain identification for outbreak investigation and clinical diagnostics.

Key Features:

• Direct ST prediction: Predicts sequence types (STs) directly from raw, uncorrected long reads without prior error correction or assembly.
• Real-time processing: Consumes sequencing data as it is generated to provide near-real-time ST predictions.
• Supported sequencing technologies: Accepts long-read data from Oxford Nanopore Technologies and Pacific Biosciences.
• Performance metrics: Demonstrated sensitivity of 94% and specificity of 97% in tests with over 700 isolates sequenced on Oxford Nanopore and Pacific Biosciences platforms.
• Speed: Delivers ST predictions on average within 90 seconds per isolate.
• Implementation: Implemented in Python.
• License: Distributed under the GNU General Public License version 3 (GPLv3).

Scientific Applications:

• Bacterial Outbreak Investigation: Rapidly determines whether isolates belong to an outbreak strain to support containment and epidemiological analysis.
• Clinical Diagnostics: Provides timely MLST-based strain characterization to inform clinical decision-making.

Methodology:

Processes uncorrected raw long reads from Oxford Nanopore Technologies and Pacific Biosciences in real time, bypassing explicit error correction and assembly steps to produce immediate MLST (ST) predictions.