Assembly

Meraculous: de novo genome assembly with short paired-end reads

Motivation

Meraculous has been developed in order to improve the quality of short-read assembly of small and large genomes.

Results

Meraculous was benchmarked against other available assemblers on a fungal dataset from Illumina as well as an E. coli dataset commonly used as a benchmark. The new assembly method is distinguished from similar methods by (1) accuracy at both the base level and on large scales (fewer mis-joins), (2) large-scale parallelization of several steps, and (3) introduction of a novel memory-efficient hash scheme for the most memory intensive step. While initially developed for JGI's plant genome program, Meraculous has been tested for microbial assembly and shown as providing the best assembly compared with other available methods by various metrics.

A paper describing Meraculous has been submitted for publication. The manuscript also describes analytical methods for the characterization of short-read datasets in the absence of a reference assembly, and several of these methods have been adopted by the QC group and others at JGI.