Scientific software is all too often cumbersome to install and execute. The work described in a journal article is backed by code in a repository, with a set of instructions to install and execute. Often, after following these instructions, one is still unable to execute the software. Missing or incompatible dependencies, ineffective interfaces or lack of documentation are some of the reasons that frustrate users and may choose to not use some software altogether. We found our software tools exhibiting some of these pitfalls, so we started a journey towards more effective distribution of our bio2Byte Tools (b2BTools). We deployed them to a web server, which allowed their execution via web interface or REST-API. This approach carries some limitations, explained in board “web solution”. In this work, we present the deployment of our tools as a python package, which can be executed locally, and circumvents the limitations that our web solution has. Our experience is explained as a use case of how we believe scientific software should be published to increase its impact and integration in pipelines.

{"references": ["Kagami, L. P., Orlando, G., Raimondi, D., Ancien, F., Dixit, B., Gavald\u00e1-Garc\u00eda, J., Ramasamy, P., Roca-Mart\u00ednez, J., Tzavella, K., & Vranken, W. (2021). b2bTools: online predictions for protein biophysical features and their conservation. Nucleic Acids Research, 1. https://doi.org/10.1093/nar/gkab425"]}

Presented in the conference "Applied Bioinformatics in Life Sciences (4th edition)" during the second poster session on March 11th 2022