Dr Stephan Heunis

Data management and training platform. A decentralized data management infrastructure will help focus on developmental and therapeutic longitudinal data, training all participating researchers in the necessary skills for future use. This strategy will lay the foundations for further data-driven computational modelling projects in the next funding period. This is a distributed project, with representatives at all main TRR379 sites. As a key software solution, this project employs DataLad. DataLad is a data management software designed to facilitate the organization, sharing, and reproducibility of scientific datasets. It integrates version control with data handling, allowing researchers to track changes, collaborate efficiently, and ensure the accessibility and integrity of their data. By leveraging tools like Git and Git- annex, DataLad provides a streamlined way to manage large datasets, making it particularly valuable in fields like neuroimaging and bioinformatics.

Over the past two years, our team has been working on an interoperable software toolstack that is open source, self-hosted, and covers basic relevant needs of a computational neuroscience lab.Notably, a number of software solutions came into existance or were deployed or further developed thanks to interactions of different software communities during RDM workshops or the distribits conference for distributed data management technologies (distribits.live).Our objective is to design an approach that allows storing data of arbitrary size, flexible semantic meta data, and the relations between these data; and to provide ways to query those relations and access the underlying data, as well as exposing selected data for websites, knowledge bases, or data catalogs.The system components are either fully compatible or integrated with the DataLad (Halchenko et al., 2021) ecosystem for data management.At the core of the stack, we have developed the following software components: