Research Data Management

Research data management (RDM) for the CRC 1436 is established under the umbrella of the RDM infrastructure that OVGU is currently implementing, with standards according to the FAIR principles and with GDPR conformity. The further development of RDM procedures at OVGU will allow it to advance its vision for convergence across disciplines, translation from fundamental research to health care, and integration of data across scales and species. To that end the CRC 1436 is already working within the context of NFDI4Bioimage (of which the LIN is a member) and NFDI4ING (for which Sanaz Mostaghim from OVGU serves on the Advisory Board).

The OVGU RDM will tackle challenges of having data sources from different comm­unities, establishing data and metadata standards, efficient data provenance and work­flows, and infrastructure and services. OVGU and CRC 1436 are working on shared metadata standards where the CRC has played a pivotal role in their development. The CRC 1436 has a dedicated person, Maximilian Günther, for data management who also serves the role of a Data Steward in the wider RDM of OVGU and will work together with the RDM Coordinator of the OVGU. Maintaining high standards of data provenance and workflow will be supported by a joint Neuroscience Data Management Hub implemented by the CBBS, the CRC 1436 and the DECODE Platform of the Cognitive Vitality research cluster.

The organ­isation and governance of the Neuroscience Data Management Hub will incorporate best practice procedures from various sources, including the DZNE Clinical Research Platform and the data management system that was developed at OVGU, DataLad, currently in use within the CBBS imaging platform together with the Brain Imaging Data Structure (BIDS) format. Dr Manuela Kuhn from the Department of Psychology and Dr Emanuele Porcu as the Data Steward of the Department of Psychology (25 % CBBS funding) are already supporting to integrate CRC needs into the Neuroscience Data Management Hub. This is supervised by Prof Dirk Oswald in the Department of Psychology. The Neuroscience Data Management Hub will connect existing projects and local institutions such as IT services, Data Integration Centres (DIZ), and platforms. OVGU and in particular FME provide an IT infrastructure that can fully support the CRC 1436. In addition to the DIZ and the FME, OVGU is constructing a new building hosting its HPC cluster and data storage. Moreover, the university cloud-service, fully funded by the state of Saxony-Anhalt, provides an extensive capacity for interregional data management enhancing collaboration with MLU and universities of applied sciences.

Data quality assurance is already coordinated by aforementioned Data Stewards and will be integrated into the Neuroscience Data Management Hub. the grafic provides an overview of the communities and institutions that the Neuroscience Data Management Hub will manage. Quality assurance is also supported by the experience and infrastructure set up by CRC 1436 PIs Düzel and Speck for curating, and rapid online QA performance of imaging data and their related experimental log data within the  National Neuroimaging Network in the DZNE (https://tinyurl.com/4su8u9ew). The iNET personnel, Dr Falk Lüsebrink and Eric Einspänner (both DZNE) will also be integrated into the Neuroscience Data Management Hub. Specifically, they will implement their newly developed automated MRI and experimental log data QA procedures for the entire CRC. Our experience in setting up multicentric SOPs procedures and the ISO 9001:2015 certified procedures of biosample, cognitive and anamnesis-based data collection implemented by the DZNE clinical research platform are already in use for data collection in the Z03 project, and through that for all human projects. Together with the extensive experience of the site in clinical research MR Imaging, documentation, imaging protocol-development/implementation, QA in multicentre studies, including multicentre task-fMRI, imaging protocol conformity with rapid feedback, scanner/hardware harmonisation, communication with vendors, centralised education/instruction of MR-personnel, ensures best possible data quality, reproducible data documentation and analysis for all CRC projects. Finally, LIN and OVGU are in close exchange and part of the NFDI-Neuro Initiative community to extend and employ current services and tools across scales and disciplines.  NFDI-Neuro is addressing the needs of the German Neuroscience community and tries to organize and sustain cross-community RDM in neuroscience. We develop processes for data management in particular also for data from animal experimentation and here foremost data from imaging and their QA with NFDI. Computer safety and data storage protection will be managed by the IT departments at the participating institutions in line with DFG requirements. With the support of the OVGU central IT facility local cloud solutions that work independently of the public internet are already implemented for safe data exchange among CRC members.

In 2019, the Data Integration Center (DIC) was established at the UMMD to create a standardized research data infrastructure. This includes a research storage cluster with a current capacity of 1.0 petabytes, which allows for centralized data backup and ensures further processing in accordance with Good Scientific Practice. Depending on performance requirements (bandwidth), data are provided directly to the virtualized evaluation environment or as BSI-compliant On Premise cloud storage in order to be optimally connected to the respective evaluation environments. By connecting local and cross-location user directory service, the generated research data can be released to the respective research group in a personalized manner and secured by authentication in the revision context. To enable inter­dis­ciplinary colla­boration with the other faculties of OVGU, the col­laboration system Confluence is made available. In addition, the DIC provides a GitLab-based repository for publishing and exchanging our source codes and “text-based research” data. These facilities promote the accessibility and archivingof research and metadata in accordance with FAIR principles.Additionally, safeguarding data for periods of 10 years or30 years for PET data is guaranteed.The German Centre for Neurodegenerative Diseases (DZNE) in Magdeburg complements this infrastructure by providing an XNAT system, specifically designed for the distribution of MRI data.This XNAT system is also used by DZNE’s MR-PET for the collection of Tau-PET imaging data, and in the next funding period of synaptic density PET imaging data, for the CRC.A Gitlab system is also provided by the DZNE. The data stored there can later be added to theGitlab of the DIZ. Metadata management is done for all CRC 1436 projects with RDMO tool under the guidance of Maximilian Günther. The CRC 1436 is already sharing data with other research institutions, for instance in the field of SuperAgeing.

The medical faculty is part of the Medical Informatic Initiative (MII) and the Network University Medicine (NUM) which both have already established extensive infrastructure and standardised regulations for research with medical data. The CRC 1436 is building on MII/NUM developments regarding processes and documents (e.g., consents, user rules), harmonised technical and semantic interoperability such as core data sets, terminologies, metadata, and interfaces. In the future, the Neuroscience Data management hub will operate in accordance with the International Neuroinformatics Coordinating Facility (INCF) and the data access, quality, and curation for observational research designs (DAQCORD) reporting system. An important milestone that we have been able to achieve in the current funding period is a standardized consent form which can be used for all participants across all CRC projects. Establishing was a major step supported by legal departments across faculties, the CRC and the Referat Forschung. Given that the administrative and legal challenges of this unified consent have been overcome, it will be straightforward to adapt this to the projects in the next funding period.

OVGU has a track record of making unique data sets reusable globally through open access. Examples are the Human Phantom 7T highest resolution in vivo brain imaging data ¹²¹, which were downloaded more than 60.000 times and the Forest Gump study of high resolution 7T brain fMRI of individuals watching the movie Forest Gump, a data set that is very extensively annotated ¹²². For the latter, an international data analysis contest led to several well received publications. We will continue this open science strategy with our new and unique imaging technology (i.e., 7T connectome). Through this, we will benefit from coordinated national and international developments and knowledge in medical research data management and digitalisation.

An important aspect of RDM is the specific training that is required in relation to the source of the data and aspects of data protection. We are meeting this challenge by placing a specific emphasis on teaching fundamental principles of data management, ethical considerations, and data protection through the teaching infrastructures regarding RDM (OVGU – eLearning Course in Moodle). RDM at OVGU is being taught in digital formats and on campus and information is disseminated in workshops and lectures across all cooperations such as with the OVGU Graduate Academy and the OVGU medical library. These include structures for the recruitment and super­vision of students in established Master, Ph.D., and MD programs (see section 4.1), compulsory lectures and e-learning modules on good scientific practice, one to one coaching events, centrally regis­tered laboratory note­books, and bio­s­ta­tistics cou­rses. For animal exp­erimentation, in accor­dance with the 3R prin­ciples, we strive for refine­ment of experimental app­ro­aches, reduction of animal numbers, and repla­cement by in vitro tech­no­logies. Compu­ta­tio­nal mo­d­e­lling and the use of common models on neu­rocognitive circuit fun­ction through the DECODE plat­form will be an im­portant overarching mea­sure to­wards achie­ving this goal.