Protection of Human Subjects

  • All users and projects making use of our computational facility must abide IRB policy with regards to conducting research that prevents potential identification of human subjects.
  • Two main policies are as follows:

  • No human identifiable experimental data can be hosted on shared dataspace.
  • No human identifiable primary data can be generated or distributed as the outcome of the analyses.

  • With respect to the human imaging data the following applies. All human imaging must be used only after removal of facial information. For data collected on CBIR, the facial information is filtered at the point of data transfer and therefore all CBIR data is free of facial information. For data collected elsewhere, please contact Sung Yu ( to discuss the removal of facial information. This can be done most easily by skull stripping your data and removing the originals.
  • With respect to human genomic data the following applies. All human genomic data obtained from CBIR data manager (Hema is imputed and stripped of raw genotype values. Raw genomic data CANNOT be hosted on the shared cluster space. Please see Hema if you need to analyze genomic data obtained outside of the CBIR. This will require UM IRB approval and re-coding of data. Hema will help with the preprocessing of the genomic data including assignment of the new IDs. In addition all directories containing genomic data have to be secured with group policy and encrypted when not in use. Please see Sung if you need help with this.