. 2020 Jun 17;9(6):giaa068. doi: 10.1093/gigascience/giaa068

Table 1:

Comparison of Watchdog with 4 other commonly used WMSs.

Category	No.	Feature	Watchdog	Galaxy	KNIME	Snakemake	Nextflow
Reusability	F1	Support for tool creation	Command-line/GUI	Command-line¹	Eclipse Wizard	No	NA
	F2	Tool documentation	XML based	XML based	XML based	YAML based²	NA
	F3	Tool reference book	Web page generator	Part of GUI	Part of GUI	Web page generator	NA
	F4	Tool versioning	Yes	Yes	Yes	Yes	NA
	F5	Sharing of tools	Repository³	ToolShed⁴	KNIME Hub⁵/NodePit^6,*	Repository⁷	NA
	F6	Sharing of workflows	Repository³	ToolShed⁴	KNIME Hub⁵/NodePit^6,*	Repository⁸	nf-core^9,*
	F7	Repurposing of workflows	XML edit/GUI	GUI	GUI	Copy Snakefile	Command-line
Reproducibility	F8	Software version logging	Yes	No	No	Yes	No
	F9	Software deployment	Execution wrappers, Conda, Docker	Conda, Docker	No	Conda, Docker	Conda, Docker
	F10	Creation of workflow report	Yes	List via history	Static description¹⁰	HTML report	HTML report
	F11	Citation export	Yes	Yes	No	No	No
Execution	F12	Resume workflow	Yes	No	Yes	Yes	Yes
	F13	Process only updated tasks	Yes	No	Yes	Yes¹¹	Yes
	F14	Process only new replicates	Yes	No	No	Yes¹¹	Yes
	F15	Detach/reattach	Yes	Yes¹²	Non-free feature¹³	Yes¹⁴	No

The selected WMSs are compared against Watchdog based on features grouped broadly into the categories reusability, reproducibility, and execution. NA: not applicable. ¹Python-based command-line program (Planemo). ²No explicit documentation of parameters but example Snakefile and wrapper source code is part of the documentation. ³[18]. ⁴[28]. ⁵[29]. ⁶[30]. ⁷[31]. ⁸[32]. ⁹[33.] ¹⁰A description that was manually created for a specific workflow can be displayed but is not dynamically created. ¹¹Flag "--list-params-changes" or "--list-input-changes" in combination with the "--forcerun" flag. ¹²Client: anytime/server: if jobs are not executed locally on the server. ¹³Non-free SGE extension or KNIME server required. ¹⁴Sending of a TERM signal stops scheduling of new jobs and waits for all running jobs to finish; Ctrl+C kills all jobs running on the local computer, while jobs running on a computing cluster continue to run. *Community project.