Proposal: Promoting an open-science research culture at the SWC/GCNU

Author(s): Joaquin Rapela, Stephen Lenzi, Adam Tyson and Spencer Wilson

Make open-science flourish at the SWC/GCNU.

Purpose

• Make open science a common practice at the SWC/GCNU.

Our vision for the practice of open data-science at the SWC/GCNU

• Every paper published by authors from the SWC/GCNU provides links to a repository containing the code used to generate the paper and, whenever possible, links to the complete data set, or a sample of it.

• The code of every laboratory is under version control in the internal SWC Github organization.

Why should we care about open science?

• The biomedical science is facing a replication crisis (Button et al., 2013). A large proportion (>80%) of recent biomedical findings cannot be replicated (Open Science Collaboration, 2015). Thus, it is essential that we provide readers with the code and data used to produce our publications, so that they can replicate them.

• It is not rare that when an SWC/GCNU member leaves his/her lab the code/protocols he/she developed are lost. By storing our code/protocols on an repository, we may overcome this problem and increase code reusability at the SWC/GCNU.

• The mere act of using a version control tool (e.g., git) helps you to generate better code (e.g., easier to maintain and understand).

What are we doing to promote open science at the SWC/GCNU?

To produce open science it is important that we know how to use open science tools like Git, Github, Python and Anaconda. We will build on previous work by past (Charly Rosseau and Maxime Rio) and current (Adam Tyson and Steve Lenzi) SWC members to expand open-science related activities at the SWC/GCNU. We offer to:

• Create an Open Science organization to centralize activities at the SWC/GCNU related to open science.

• Reinstate PyClub meetings on a monthly basis and PyStarter courses on a yearly basis, which have been suspended due to the COVID-19 pandemic.

• Improve the current PyClub/PyStarters web site providing:

  • Contact information of the steering committee and contributors to the organization,

  • Technical information about open data-science tools (e.g., links to good tutorials, recommended tools),

  • Information about future and past events (e.g., PyClubs and PyStarters, software carpentry courses at UCL).

• Organize advanced courses on, for example, data acquisition, image analysis, time-series analysis, statistical neuroscience and statistical models for neuronal populations.

• Setup a mentoring scheme where people with no experience in open data-science tools are paired with people with substantial experience on them.

Who will benefit by the adoption of open-science practices at the SWC/GCNU?

1. Science
   • we will deliver publications facilitating the essential scientific feature of replicability
2. Society
   • help overcome the distrust in science due to the lack replicability in the biomedical sciences
3. SWC/GCNU
   • reproducible research has a larger impact?
   • using version control increases the quality of code
   • publishing papers with code and data may motivate scientists to be more careful about their research
4. SWC/GCNU members
   • experience on open-science tools is a valuable skill to develop

How could SWC/GCNU faculty help promote open science?

1. Asking their lab members to put their code under version control (e.g., git/Github) and learn how to use it.

2. Encouraging their lab members to publish their papers with links to their code and the data (whenever possible).

3. It would help if every laboratory at the SWC/GCNU has at least one member well trained on the use of open data-science tools. This person will then help his/her lab colleagues use these tools. If a laboratory does not have such a person, the principal investigator could ask one of his/her team members to attend PyStarters.

Next steps

Refer to the Promoting Open Science project.

Stakeholders

• Any SWC/GCNU member producing scientific results.