Note: this article is the opinion of the author, Mark Ziemann, and doesn’t necessarily reflect the official position of RN4CB.
Together with 11 researchers across 7 countries, I’m ecstatic to announce the founding of the Reproducibility Network for Computational Biology (RN4CB).
I guess you’re asking why we need such a network, and why now?
Computational biology, like any branch of computational research should be 100% reproducible. The computers we use are exquisitely precise and the analyses we do should in theory be completely reproducible.
In practice however, the picture is not so straightforward. Simple things like methodological descriptions, availability of code and raw data are the main reasons why computational biology research doesn’t reproduce and unfortunately, past studies like (Ioannidis et al, 2009) have shown that just ~11% of published studies are fully reproducible.
In addition to these major causes, there are less well known ones such as the brittle nature of software packages which are difficult or sometimes impossible to reproduce just a few years after publication (Perkel, 2020). Generative Artificial Intelligence (AI) also poses a threat through the inadvertent incorporation of errors.
It is my opinion that reproducibility in bioinformatics and computational biology has been neglected for a long time, and we lack the solid data to say whether the situation has improved at all since 2009.
Research that is reproducible can be audited, so we can check for underlying problems like invalid methods or inconsistencies with methodological descriptions. These hidden problems are not easily identified in the current peer-review system.
Collective action is required at all levels from student researchers all the way to institutes, funders and publishers to improve the quality of computational biology research practices.
Having seen the excellent work done by AIMOS, COS, UKRN and Aus-RN, we saw the potential for a reproducibility network to serve the global community of life scientists. This is where RN4CB comes in.
Core mission
The Reproducibility Network for Computational Biology (RN4CB) is a peer-led, globally inclusive consortium serving as a forum and voice to advocate, educate, and set standards for open and reproducible research practices.
Our aim is to raise the quality and trustworthiness of research in computational biology, medical informatics, and related data science fields.
Aims and scope
I’m hoping that RN4CB addresses these problems with all available approaches including education, meta-research, recommendations and policy initiatives. A key role of the network is to raise awareness to these issues to various stakeholders including the research community, policy makers and the general public. The network aims to be as inclusive as possible, encompassing the fields of computational biology, biomedical informatics, cheminformatics, neuroinformatics and data-intensive medical research and psychology.
We want to include researchers and practitioners at all career levels, from undergraduate students, doctoral students, research assistants, post-docs and senior scientists in the research and industry sectors. As the reproducibility problems we face are caused by a multitude of factors, the network will involve experts in training and education, publishing, research integrity, computational research practices, domain expertise, meta-research, policy and funding. The network will work closely with related reproducibility groups and organisations representing computational biology and related fields.
Website
RN4CB has a brand new, official webpage.
Activities
The network will be involved in the following activities to further the mission:
- Support education through workshops and tutorials
- Events (journal club, seminar, online symposium, etc)
- Developing best practice guidelines, recommendations and policy
- Media and online presence incl newsletter, social media and website
- Funding initiatives
- Awards for excellence
- Career development opportunities
- Cooperation with other bodies
Plan for the rest of 2026
From September, the committee will be hosting events and running a membership drive. Later in the year, the committee will elect a Chair, a role that I am assuming in the interim on a caretaker basis. Stay tuned for more updates.
References
Ioannidis JP, Allison DB, Ball CA, et al. Repeatability of published microarray gene expression analyses. Nat Genet. 2009 Feb;41(2):149-55. doi: 10.1038/ng.295. Epub 2008 Jan 28. PMID: 19174838.
Perkel JM. Challenge to scientists: does your ten-year-old code still run? Nature. 2020 Aug;584(7822):656-658. doi: 10.1038/d41586-020-02462-7. PMID: 32839567.