A graph-based approach for designing extensible pipelines

BMC Bioinformatics

Table 2 Executable pipelines for three usage scenarios

File	Code
SDAT to R HierFstat
.inputs	SDAT02
	SDAT202
.exec	perl SDAT2Rhierfstat.pl SDAT02 SDAT202 RHierfstat02
.outputs	RHierfstat02
PolyPhred output to Structure
.inputs	PolyOut01
.exec	perl PolyPhred2PrettyBase.pl PolyOut01 PrettyBase01
	perl PrettyBase2SDAT.pl PrettyBase01 SDAT01
	perl SDAT2Structure.pl SDAT01 StructureFormat01 mainpar01 extrapar01
.outputs	StructureFormat01
	mainparamt01
	extraparam01
Phase to Fasta
.inputs	PHASEOUT03
	Fragments03
	RefSeq03
.exec	perl Phase2Fasta.pl PhaseOut03 Fragments03 RefSeq03 Fasta03
.outputs	Fasta03

Executable pipelines for file-format conversions: (top) SDAT format to R Hierfstat input format; (centre) PolyPhred output format to Structure input format; and (bottom) software PHASE output format to Fasta format. In practice, input and output files handled by the pipeline system are renamed to include a timestamp identifier of each specific pipeline (such as numbers 01, 02 and 03 above). This guarantees that inputs and outputs stored in the system are unique for each dynamically generated pipeline.

ISSN: 1471-2105