Watchdog 2.0: New developments for reusability, reproducibility, and workflow execution

Abstract

BACKGROUND: Advances in high-throughput methods have brought new challenges for biological data analysis, often requiring many interdependent steps applied to a large number of samples. To address this challenge, workflow management systems, such as Watchdog, have been developed to support scientists in the (semi-)automated execution of large analysis workflows. IMPLEMENTATION: Here, we present Watchdog 2.0, which implements new developments for module creation, reusability, and documentation and for reproducibility of analyses and workflow execution. Developments include a graphical user interface for semi-automatic module creation from software help pages, sharing repositories for modules and workflows, and a standardized module documentation format. The latter allows generation of a customized reference book of public and user-specific modules. Furthermore, extensive logging of workflow execution, module and software versions, and explicit support for package managers and container virtualization now ensures reproducibility of results. A step-by-step analysis protocol generated from the log file may, e.g., serve as a draft of a manuscript methods section. Finally, 2 new execution modes were implemented. One allows resuming workflow execution after interruption or modification without rerunning successfully executed tasks not affected by changes. The second one allows detaching and reattaching to workflow execution on a local computer while tasks continue running on computer clusters. CONCLUSIONS: Watchdog 2.0 provides several new developments that we believe to be of benefit for large-scale bioinformatics analysis and that are not completely covered by other competing workflow management systems. The software itself, module and workflow repositories, and comprehensive documentation are freely available at https://www.bio.ifi.lmu.de/watchdog.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 11 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.