Collaborative software development to disentangle counterproductive incentives

Abstract

Background: We propose a collaborative approach to address the conflicting incentives between software development (generating high-quality code) and research (generating scientifically interpretable results). To this end, the team of the Genome Competence Center at the Robert Koch Institute (RKI) collaborates with the computer science programs at HTW University of Applied Sciences.
Approach: While bioinformaticians at RKI have to focus on analysis quality and speed, the partners at HTW can concentrate on software robustness and maintainability. Challenges include access to real data and the time restrictions of student projects. Agile software development methods such as frequent joint meetings, the use of the collaboration tools provided by Azure DevOps and hackathons in addition to a long-term project coordinator are employed to alleviate these problems.
Conclusion: This approach has led to the development of a new quality control pipeline following software engineering best practices: Coding conventions (e.g. adherence to PEP 8 for Python code), an extensive test suite, version control with branching guidelines and gated commits, and automated reproducible builds. Our experiences highlight the benefits of such collaborations, and we hope to encourage others to follow similar approaches in order to facilitate the development of high-quality research software. The newly developed QC pipeline is available under https://dev.azure.com/RKI-HTW/NGS-QC/_git/QCurchin