BASE: a practical de novo assembler for large genomes using long NGS reads.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• De novo genome assembly from next-generation sequencing (NGS) data is computationally intensive, particularly for large genomes.
• Existing assemblers like SOAPdenovo2 (de Bruijn graph-based) are efficient but do not fully leverage longer NGS reads.
• String graph assemblers (e.g., SGA) handle longer reads better but are generally slower.

Purpose of the Study:

• Introduce BASE, a novel de novo genome assembler designed to improve efficiency and contig quality for NGS data.
• Address the limitations of current assemblers in utilizing longer NGS reads effectively.
• Provide a practically efficient tool for genome assembly, with potential for scaffolding extension.

Main Methods:

• BASE employs an enhanced seed-extension approach, utilizing efficient read indexing to generate adaptive seeds.
• High-probability unique seeds are used to construct extension trees.
• Reverse validation, incorporating read coverage and paired-end information, refines branches to produce high-quality consensus sequences for contig extension.

Main Results:

• BASE demonstrates superior contig quality and speed compared to SOAPdenovo2 and SGA for longer NGS reads (up to 250 bp).
• On bacteria datasets, BASE achieved better quality than SOAPdenovo2 and SGA, comparable to SPAdes for 250 bp reads.
• BASE showed higher N50 values across human datasets, with significant improvements at 250 bp read lengths, and exhibited better memory efficiency than SOAPdenovo2 with error-prone data.

Conclusions:

• BASE is a practical and efficient tool for de novo genome assembly, offering significant quality improvements for long NGS reads.
• The assembler's design facilitates straightforward extension to include scaffolding capabilities.
• BASE represents a valuable advancement in bioinformatics tools for genome assembly research.