DIIL Policies

Project Form: The website provides a project form that investigators fill out to describe what resources they need. CIVM staff will be happy to assist with any questions regarding the form. The crucial information included in the document helps us understand what you need and when you need it. Regulatory (IACUCC) and billing contacts are crucial. Estimates of animal numbers and time frames for the study help us balance the load on our resources. A project cannot start until the form is completely filled out. A code will be issued eg. year.PI_name.project key. An example, 20.JDoe.03 tells us that JDoe started his third project in 2020. Scheduling, data archive and billing are keyed to this code. A project form will be active for one year from the start date. Projects must be updated annually in order to keep a clear idea of the resource use. Submission of the form constitutes acceptance of all policies at CIVM.

Acknowledgement: Papers and meeting abstracts resulting from work supported by the CIVM must acknowledge the use of the facilities. The PI must provide the Director a final copy of any manuscript or abstract at least three weeks prior to submission for publication or conference inclusion. DIIL is the collaborative arm of CIVM so there is a presumption of inclusion of faculty/staff on any of the publications using DILL resources and intellectual property. The Director will determine who at DILL has contributed to the work.

Scheduling: Once a project is approved, PI’s can schedule the resources they need via an email request. Each resource has unique constraints. PIs are urged to discuss their needs directly with the scheduler for that resource. Requests should be made as follows:

GRID servers: James Cook, jjc29@duke.edu
Use of the Big (Imaging) Data resource requires inclusion of that resource in the project form and approval of the Director as this resource is currently very heavily used by members of DIIL. Some of the packages on the GRID servers have single licenses so only one user can be operational at any given time. We will set up a weekly calendar to coordinate activity. Provide your desired times and duration and we will do our best to accommodate you.

Agilent 9.4T MRI: G Allan Johnson, gjohnson@duke.edu
The Mouse Connectome Scanner (9.4T Agilent) is the most heavily used system at CIVM. Thirty percent of the available time has been set aside to provide users who are not members of DIIL access to this unique resource using protocols that DIIL faculty and staff develop for those users. The complex integration of this system with cluster and pipelines, typically requires 3 members of the DIIL staff 6-8 weeks for construction and validation of new protocols. These protocols will be designed to work in 7 hour blocks to allow acquisition of two specimens per day (typically 9AM-4PM, 5PM-12PM). Users requesting time should send an email to gjohnson@duke.edu at least 30 days prior to their need. Time will be schedule in blocks of up to one week at a time to accommodate group studies.

Data Provenance and Backup: Data acquired on the Mouse Connectome Scanner will be archived in the CIVM Oracle data base either as raw (short integer) or .nii format. Compressed sensing data (magnitude images reconstructed on the cluster) will be archived on the CIVM Oracle data base in raw (short integer) format under the project code. Conventional (Fourier Transform) data will be stored in a similar fashion. Post processed data will be stored in .nii format. Since the data types are 3,4, and 5 dimensional, users should contact James Cook for details of data formats and data structure. We will work with the PI or his/her designee to establish accounts and test downloads. If assistance is needed e.g. downloading or using open source imaging packages, that support will be provided on an as needed basis at the standard rate ($65/hr.). Users are ultimately responsible for verifying that their data has been appropriately archived. They must review their data (from the archive) within 30 days of the study to verify there are no problems. Data will be purged from the cluster and pipelines after 30 days.