Skip to content

forbow-lab/neuro-raw-pipeline

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 

Repository files navigation

Neuro Raw Pipeline

Raw data / DICOM ingestion and organization

1. Download DICOMS (T1w/T2w/DWI)

  • Prior to leaving 3T MRI request the MRI Technologist "push" the exam DICOMS to the PBIL-master Node.
  • From an iMac in the PBIL, download the DICOMS from PBIL-master onto the Shared drive:
    • Open Finder window, create a new subject folder on shared drive (e.g. /shared/uher/FORBOW/rawdata/100_C_20190320/)
    • Open Horos
      • Select "Query" on main interface
      • Search "Forbow", select date-range (e.g. "last 7 days")
      • Select "Query" from middle of query window, within seconds a list of all Forbow scans in last 7 days should appear.
      • Download new exams by hitting the green download button beside each FORBOW participant
      • Monitor the "Activity" panel on the left side of the main Horos window until all downloads are complete.
      • Before closing the query box, ensure that no files are still transferring in the "Activity" panel.
      • From main Horos window, select four main sequences [T1w, T2w, DTI-30DIR, DTI_B0], drag into Finder window to drop on subject folder.
      • Monitor the "Activity" panel again until all DICOMS are exported to subject folder on shared drive.
      • Rename and organize the exported DICOM folder under subject folder (e.g. /shared/uher/FORBOW/rawdata/100_C_20190320/DICOMS/).
      • Note, if there are multiple T1w_BRAVO and/or T2w_FlairPrep sequences, only copy over the highest quality scan.

2. Data Quality ratings and Reminders

  • Create a "readme" file that contains pertinent notes about that particular participant and scan (e.g. T1 BRAVO collected twice due to motion, kept 2nd ) as well as the exam number
  • Create a file to rate the quality of the T1 BRAVO, T2 CUBE, and T2 PROMO on a four point scale (e.g. 109_A_qc_ratings). See any previous rawdata folder for an example
  • Complete the remaining fields of the neurolog
  • Send images to participants that requested a picture of their brain

3. Download Pfiles (RS)

  • First. confirm Pfiles are ready to download from remote BIOTIC server:
  • ssh biotic@lauterbur (enter password)
  • Navigate to data folder on remote server cd /biotic/data/Forbow/
  • ls to see the list of participant folders
  • ls 100_C to list contents of specific participant folder
  • If participant files appear with timestamps they are ready to download, if not, you'll have to wait for BIOTIC to convert them.
  • exit to close remote connection.
  • Once Pfiles are ready, download from remote BIOTIC server:
  • open terminal and cd into SUBJECT ID (e.g., /shared/uher/FORBOW/rawdata/100_C_20190320/
  • use rsync command below (all one line) replacing IDs with the correct subject IDs and date:
  • rsync -av biotic@lauterbur:/biotic/data/Forbow/100_C/ /shared/uher/FORBOW/rawdata/100_C_20190320/RS/
    • this copies RS Pfiles from remote server to our local server into the participant's RS folder.
    • make sure to supply correct subject ID and scan label for ###_X and the correct date of scan for YYYYMMDD
    • enter password if requested, watch terminal to monitor file transfer until complete and returned to prompt.

Note: may take 5-10 minutes to copy across network (~17 GB)


4. Run Rawdata Pipeline Script to convert, deface, and QC.

  • The Rawdata Pipeline Script:

    1. converts RS Pfiles into NIFTI format (1-2 hrs per subject),
    2. DICOM sequences into NIFTI,
    3. structurally defaces the anatomicals,
    4. generates a simple QC HTML page for ensuring sequences are ready for analysis,
    5. creates BIDS folder for functional processing.
  • From McCoy/Picard iMac:

    • Open Terminal
    • remote login to Jaylah: ssh uher@jaylah
    • navigate to rawdata folder: cd /shared/uher/FORBOW/rawdata/
    • start pipeline script: nohup ./_1_scripts/FORBOW_0_Run_RAWDATA_Pipeline.sh 101_C >nohup_rawpipe_100_C_20190411.txt 2>&1 &
    • monitor script is running with ps command: ps aux | grep '100_C' or htop command ('q' to quit htop)
    • pipeline takes 2-3 hrs to complete (automatically uses 4-8 CPU Cores on Jaylah)
    • from Finder window, check subject _QC_report.html to confirm completeness, eg: open /shared/uher/FORBOW/rawdata/100_C_20190320/NIFTIS/QC/_QC_report.html

Note, the Rawdata Pipeline script runs each of the below scripts, in order:

  1. Convert raw Pfiles 1_run_mux2nii.sh
  2. Convert DICOMS into NIFTIS 2_create_niftis.sh

About

Raw data ingestion and organization

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published