DNACLUST

DNACLUST clusters millions of short DNA sequences to group highly similar sequences for taxonomic profiling and phylogenetic analysis of microbial communities using markers such as 16S rRNA.

Key Features:

• Greedy clustering: Employs a greedy clustering strategy that ensures the radius of each cluster does not exceed a user-specified similarity threshold.
• K-mer based filtering: Uses k-mer based filtering to reduce the number of sequence comparisons and improve computational efficiency.
• Sequence alignment algorithms: Integrates sequence alignment methods to evaluate similarity within the clustering process.
• Scalability: Designed to process millions of short DNA sequences for large-scale dataset reduction and analysis.
• Performance comparison: Operates approximately an order of magnitude faster than CD-HIT in exact mode and matches the speed of UCLUST in approximate mode, with efficiency gains at higher similarity thresholds for removing duplicates and near-duplicates.
• Multiple sequence alignments: Can generate multiple sequence alignments for each cluster to support inspection and downstream analyses.

Scientific Applications:

• OTU identification: Supports operational taxonomic unit (OTU) identification for microbial community profiling using phylogenetic markers such as 16S rRNA.
• Taxonomic profiling: Facilitates taxonomic profiling of microbial communities by providing mathematically defined clusters that represent groups of similar sequences.
• Phylogenetic analyses: Reduces dataset sizes for downstream phylogenetic analyses and assists in estimating an environment's phylogenetic composition.
• Dereplication: Enables rapid elimination of duplicates and near-duplicates from sequencing datasets.

Methodology:

Greedy clustering with a user-specified similarity radius combined with k-mer based filtering and sequence alignment; clusters can be output with multiple sequence alignments.