e-Science Requirements

With the establishment of nationally-funded imaging equipment like 7T MRI, MEG and integrated PET/MRI, which will be used by researchers na-tionwide, the need for data handling, storage, post-processing and analysis will significantly increase. Therefore, there will be a demand for solutions for integration and transfer of image data between sites, standard postpro-cessing methods and data analysis, storage of raw and post-processed data, cost calculations, jurisdiction and rules for storage versus archiving. As the overall need can only be estimated, dynamic secure storage capacity which is easily accessible is needed. From a researcher´s perspective prerequisites for well-functioning national facilities are fast on-time access, qualified post-processing and data analysis at national servers with rapid feed-back functions to the researcher at a local level. Here, one possibility would be to use, the SweStore server hardware of SNIC and dedicated user interfaces (web portals), suitable for data with a wide range of open and proprietary image formats. A few examples of the development are:

• The National 7T MRI in Lund will begin to produce significant volumes of imaging data as of 2015 estimated to rise to 650 TB by 2020.

• Whole-body integrated PET/MRI in Uppsala will produce images that will primarily be stored at the local research PACS server but national data storage space will be needed estimated to reach 250 TB by 2020.

• The National MEG facility in Stockholm will produce 1 TB per research study and therefore, depending on the number of research studies/year, be estimated to require at least 50TB by 2020.

• Super resolution and 4D microscopy center (UMU) will need at least 50 TB/year by 2020 at a national level, in addition to the 16 TB required at the local level.

• U-CAN project, a joint cancer project between Uppsala and Umeå Uni-versities that will need storage space for data in DICOM as well as non -DICOM formats on national services of 10 TB in 2014 with an increase to 25 TB by 2020.

• Large national study (SCAPIS) funded by the Heart-Lung Foundation will need archive and storage space for large volumes of imaging data, in overall estimated to at least 100 TB/year.

• SciLifeLab in Stockholm, with 2 new STED microscopes, is expected to need 5 TB/year so in total 40 TB to 2020.

• CARS and super resolution fluorescence microscopy in Gothenburg is ex-pected to need 6 TB/year, totalling 48 TB to 2020.

• Overall estimations for need of storage space for microscopy data is at least 180 TB by 2020 (starting 2014 with 85 TB and estimated increase with 15 TB/year). The need for storage space for medical imaging (CT, PET, MRI, MEG etc.) will be at least 1000 TB by 2020 (starting with 300 TB in 2014 and estimated increase of 80-100 TB/year).

A limited number of standardized post-processing methods for analyses of medical images need to be provided, enabling the submission of imaging data by individual researchers or a group of researchers. These data, after post-processing and/or analysis, should be sent back to the researcher/

group. There is a need for national consensus in the research community for which standardized post-processing methods should be provided. The pos-sibility to “hook-on” individual post-processing software for analysis of data will be required. Solutions for handling patient/subject information need to be provided.

The access to and transport of imaging data needs to be secure, fast, and constantly maintained. There will be a need for service in the form of soft-ware developers, developers of medical imaging tools and post-processing programs and for implementation, security management and administra-tion both of the SweStore hardware servers and the different hubs. The pos-sibility to add national or individual hubs and web portals and increase the data storage space has to be a dynamic process. Currently, a group within Lunarc, at Lund University, is doing pilot studies of different possible

sce-narios for such implementation for the national 7T projects/ LBIC (Lund Bioimaging Centre). Finally, the costs for the individual researchers need to be kept low to encourage high demand and use.

There are currently several limitations that need to be addressed on a na-tional level, legal as well as ethical and issues related to maintenance of data servers. For example (i) the proposed changes in laws regarding the use of social security numbers to identify subjects and regarding data base regis-ters, (ii) the need for national regulations regarding the right to access data through local firewalls, (iii) the ownership of acquired data, and (iv) the increasing demand for open access. Other issues are (v) access time windows for short term (storage) and long term (archiving) data access as well as ar-chiving responsibilities.

In the near future, there will be an increasing need for an infrastructure related to tracers for PET and microscopy. There are also new imaging mo-dalities coming on the market that might, in the reasonably near future, be of national interest and a need for national imaging modalities, such as phase contrast CT, Electron Paramagnetic Resonance Imaging (ERPI), Mag-netic Particle Imaging (MPI), micro-MRT might require additional national infrastructure support.

8.5 Molecular Medicine

Molecular medicine is a broad field, where physical, chemical, biological and medical techniques are used to identify fundamental molecular and genetic errors leading to disease, and to develop molecular interventions to correct them. To merge molecular patient data with clinical information is believed to be one of the keys to successful medical research in Swe-den. Our advantages include our unique personal identification numbers, a long history of keeping patient records, national quality registers and our biobanks. In Sweden it is possible to perform prospective longitudi-nal studies which are of great value to research. Internatiolongitudi-nally and na-tionally there are efforts to include next generation sequencing and other large scale analysis techniques into healthcare and diagnostics. The results should form the basis for new treatment strategies. A common problem is that the local health care system fails to offer sufficiently large cohorts of patients with unusual diseases to generate statistical relevance in clinical studies. There is then a need for a well-functioning system for collabora-tion at the nacollabora-tional or sometimes at the internacollabora-tional level. A good model is haematologic malignancies, which are heterogenic and each disease

vari-ant is rare. In total there are about 2,000 new cases every year. Samples from the different sub-types are biobanked nationally and form the basis for reaching clinical relevance. The Swedish government has made a large effort to improve the national quality registers by increasing the financial support step-wise during coming years. There has also been a commit-ment to build a national biobank infrastructure financed by the govern-ment through VR. Also many of Swedish county councils are committed to develop the biobanking infrastructure locally especially at the univer-sity hospitals with national collaboration through the national biobank council.

The clinical users are often in need of bioinformatics competence and computing skills. It is therefore important that there exist wellfunctioning service centres that can provide support to the clinical scientists.