CDKAM

CDKAM classifies metagenomic sequences from third-generation sequencing (TGS) data using discriminative k-mers and an approximate matching algorithm to achieve species-level taxonomic identification despite high TGS error rates.

Key Features:

• Discriminative k-mers: Uses discriminative k-mers to represent taxon-specific sequence signatures for metagenomic classification.
• Approximate matching strategy: Employs an approximate matching algorithm to accommodate the high error rates and error patterns characteristic of third-generation sequencing (TGS) when identifying k-mers.
• Two-phase approach: Implements a rapid quick mapping phase to identify candidate k-mer matches followed by a dynamic programming refinement phase to improve match accuracy.
• Performance characteristics: Evaluated on simulated and real TGS datasets, CDKAM outperforms existing methods for sequences ~1000–1500 bases, achieving higher species-level accuracy with reduced memory use.

Scientific Applications:

• Metagenomic taxonomic classification of TGS data: Classifies metagenome sequences generated by third-generation sequencing technologies to enable taxonomic profiling.
• Species-level identification in microbial community studies: Provides high species-level accuracy useful for analyses of complex microbial communities.

Methodology:

CDKAM applies discriminative k-mers with an approximate matching algorithm implemented as a two-phase process consisting of a quick mapping phase followed by dynamic programming refinement.