Skip to content

Latest commit

 

History

History
29 lines (27 loc) · 3.65 KB

WorkflowOverview.md

File metadata and controls

29 lines (27 loc) · 3.65 KB

5.4. Workflow Overview

As noted above, modules outside of segmaps and transferanno can be launched and allowed to run continuously. These steps outline how a slide might process through the workflow.

  1. Update the AstroPath files in the <mpath>
  2. Update the slides in the Specimen_Table_N.xlsx
  3. Scan the slides according to the documentation in 4.4.
  4. Launch the AstroIDGen module to intialize the slides into the pipeline
  5. Launch the TransferDeamon module to start transferring completed slides from the <Spath> location to the <Dpath>\<Dname>
  6. Launch the meanimages module to create mean images of each slide as it finishes transferring; these will be used to build the batch flatfields. This is usually launched after all slides from a batch have been transferred as it only loops through the Projects once then stops. Creating the meanimage for all slides in a batch is also the trigger for the next step, but the code does not yet have the ability to know beforehand how many slides are in each batch which may cause a false start of the next step.
  7. Launch the flatw_queue module to create the flatfields and assign new slides to the flatw_queue. Then the module distributes flatw jobs to the assigned workers.
  8. Launch the flatw_worker module on the assigned worker machine to process the flatfielding and image warping corrections on a particular slide's hpf image set
  9. Create the BatchID and MergeConfig files for the project according to documentation in scans
  10. Launch the mergeloop module to initialize necessary antibody processing folders, create the local inform_queue for a project, recieve inform results, merge the data and create qa qc images to evaluate the inform classification algorithms.
  11. Create phenotype algorithms in inForm according to the protocol established in 5.8.3.
  12. Launch the inform_queue module to send jobs from the main inform queue to the queues on the inform worker machines.
  13. Launch the inform_worker module to process algorithms in inForm
  14. wait for QC images to process by the mergeloop module
    • evaluate the qc by the protocols established in 5.8.6. -repeat 11-13 as needed
  15. Launch segmaps module after qc has been completed for the cell classification of slides in a project to build the final segmenation maps.
  16. Launch transferanno module after the slides have been successfully annotated in HALO and annotations have been exported to a desired location.
  17. Launch cleanup and the associated script convert_batch from the cleanup module
  18. manually unmix the control tma component tiffs and place them in the inform_data\Component_Tiffs folder of each respective tma
  • AstroIDGen should be launched on the <Spath> location.
  • The following modules can be launched at the same time: TransferDeamon, meanimages, flatw_queue, inform_queue, mergeloop.
  • The following modules must be launched on their respective "worker" locations: flatw_worker and inform_worker.
  • Launching segmaps runs over the entire set of projects but is not a continous loop and must be restarted to reprocess
  • transferanno is a simple script that distributes and renames the halo annotations from a single folder to corresponding slide folders in a single project folder