NanoCLUST

NanoCLUST classifies amplicon-based full-length 16S rRNA nanopore reads through unsupervised clustering and consensus building to enable species-level bacterial identification and abundance estimation.

Key Features:

• Unsupervised Read Clustering: Clusters reads using Uniform Manifold Approximation and Projection (UMAP).
• Consensus Construction: Builds a polished read from clustered sequences.
• Taxonomic Classification: Performs Blast classification of the polished consensus read.
• Species-Level Resolution: Enables accurate bacterial identification and abundance profile estimation at species level.

Scientific Applications:

• Mock Community Characterization: Characterizes commercial mock communities with higher accuracy than other state-of-the-art software.
• Bacterial Identification: Identifies bacteria from full-length 16S rRNA nanopore sequencing data.
• Abundance Profiling: Estimates bacterial abundance profiles at species-level resolution.

Methodology:

NanoCLUST uses an unsupervised read clustering step based on Uniform Manifold Approximation and Projection (UMAP), followed by construction of a polished read and subsequent Blast classification.