SLICEMBLER

SLICEMBLER performs de novo genome assembly of ultra-deep sequencing data (coverage ≥1000x) by partitioning reads into slices, assembling each slice with standard assemblers, and integrating assembly results into a consensus of accurate contigs.

Key Features:

• Data Partitioning: Partitions input sequencing data into optimally sized "slices" to enable independent assembly of ultra-deep (≥1000x) datasets.
• Slice Assembly with Standard Assemblers: Assembles each slice independently using Velvet, SPAdes, IDBA_UD, and Ray.
• Consensus Assembly: Applies a majority voting mechanism among individual slice assemblies to identify long contigs for integration into a consensus assembly.
• Efficiency Optimization: Utilizes a generalized suffix tree to detect frequent contigs or contig fragments across different assemblies and aid accurate merging.
• Error Resistance: Experimental results report that SLICEMBLER outperforms base assemblers in error resistance, producing substantially fewer assembly errors and in some cases error-free assemblies under high sequencing error rates.

Scientific Applications:

• Ultra-deep sequencing projects: Assembly of genomes from ultra-deep sequencing datasets with coverage levels of 1000x or higher.
• Complex and repetitive genomes: Precise assembly of complex genomes and highly repetitive regions that require resolution of long contigs.
• Error-prone sequencing contexts: Generation of robust, error-resistant assemblies for datasets with high sequencing error rates.

Methodology:

Partition reads into "slices"; assemble each slice with Velvet, SPAdes, IDBA_UD, or Ray; perform majority voting across individual assemblies to identify long contigs; use a generalized suffix tree to detect frequent contigs or fragments and merge them into a consensus assembly.