Neurodesk is a community project providing a containerised data analysis environment to facilitate reproducible analysis of neuroimaging data.
About
Neurodesk is a data analytics computing environment designed for flexibility, scalability and reproducibility, particularly in the field of neuroimaging. In a nutshell, Neurodesk provides a containerised data analysis environment where users can build and use containers to analyse neuroimaging data. About 300 GB of neuroimaging software packages are stored in CVMFS repositories replicated across Australia, the United States and Europe, serving around 300 European users per month.
The Challenge
Neuroimaging research often involves a complex analysis pipeline that relies on specific software releases and operating systems. This creates challenges in reproducibility, as running the same pipeline on a different computer can lead to unexpected results.
Neurodesk addresses this issue by providing a containerised data analysis environment. This environment ensures that researchers can access the exact software versions and tools they need, regardless of their computing setup. This allows for:
- Reproducible Research: Findings can be easily replicated by other researchers, promoting trust and accelerating scientific progress.
- Simplified Collaboration: Researchers can share their analysis pipelines and data with collaborators without worrying about compatibility issues.
- Increased Efficiency: Scientists can focus on their research questions rather than troubleshooting software dependencies.
The Neurodesk project requires two computational resources to function well in a particular geographical region: a CVMFS Stratum 1 mirror and a so-called Play server. The project has used custom CVMFS mirrors on public cloud providers in the past, but managing such a server requires dedicated people to keep everything updated and running optimally. EGI provides access to large shared CVMFS mirror servers that are professionally managed distributing the Neurodesk repository through this infrastructure makes use of existing and well-managed infrastructure and increases the resilience of the network. The Play server is a cloud-hosted Neurodesk environment that allows potential Neurodesk users to explore the software without installing anything locally. It is also often used in workshops where researchers need to access a consistent computational environment. Providing cloud resources for this Play service in Europe allows European researchers to access this environment quickly and efficiently.
A screenshot from running a computational neuroimaging notebook in Neurodesk Play
Neurodesk uses these services from EGI
- The EGI Cloud Compute and the cloud-based EGI Online Storage to host the Neurodesk Play server in the EGI Federation.
- The EGI Check-in to enable secure access to the EGI resources pool used for hosting the Neurodesk Play server.
- The EGI Software Distribution to deploy up to 1TB of neuroimaging datasets on a worldwide distributed computing infrastructure.
- The EGI training and consultancy service
The Neurodesk data analytics computing environment is supported by the following partners
with funding from
Impact
> 300
European users served/month
300 GB
of neuroimaging software hosted in the EGI Stratum 1 (1 TB foreseen)
4,565 Cloud CPU/H
consumed since June 2024
The partnership with EGI allows European users to access the Neurodesk project through high-quality caching and cloud infrastructure. With this, European researchers can now access a wide range of Neuroimaging software on European computing resources and process Neuroimaging data in a reproducible environment.
Steffen Bollmann (The University of Queensland)
