doepipeline: a systematic approach to optimizing multi-level and multi-step data processing workflows

BMC Bioinformatics

Table 1 Factors in the de-novo assembly case

Parameter	Abbr.	Type	Min	Max	Default	Optimized
Size of k-mer (k)	KMER	Ordinal	20	90	31^a	38
Minimum mean k-mer coverage of a unitig (c)	MIKC	Quantitative	2	15	sqrt (median)^b	8.5
Minimum alignment length of a read (l)	MIAL	Ordinal	20	60	40	30
Minimum number of pairs required for building contigs (n)	MIPA	Ordinal	5	15	10	15

The four factors investigated in the de-novo assembly case are described above. The letter in parenthesis following the parameter name is the parameter used in the abyss-pe command line interface. Min and max values define the design space. ^a: There is no default value explicitly specified by the ABySS documentation. However here we used a k-mer size of 31 for comparison purposes. ^b: This refers to the square root of the median k-mer coverage, which is affected by the sequencing depth and choice of k-mer size. The optimized values are the combination of factor values that produced the best outcome, as found by doepipeline

ISSN: 1471-2105