basic-uq-hpc-scripts

A collection of scripts for processing calcium imaging data using the University of Queenslands HPC infrastructure.

Main steps in processing data

• Connecting to / using HPC
• Run Suite2p
• Run ANTs

Initial installation

• For now I will assume ANTs and Suite2p are installed already
• To initially install this repository log on the HPC (described below) then run

git clone https://github.com/Scott-Lab-QBI/basic-uq-hpc-scripts.git

A few one of things to change

• Update line 5 of suite2p_hpc.pbs with your school name (it is likely one of: UQ-QBI, UQ-SCI-SBMS or UQ-EAIT-ITEE), use the groups command to see which group you belong to.

Connecting to the HPC

• Open Acaconda (or Powershell) prompt
• Type (or copy and paste) the following into anaconda prompt

ssh <uqusername>@tinaroo.rcc.uq.edu.au

• Replace <uqusername> with your uq username (e.g. uqjsmith or s4323534)
• If Bunya has been installed, replace the link to tinaroo.rcc.uq.edu.au with the new link to bunya.
• Alternatively, you can use VSCode, wsl, Putty or any other ssh client instead of anaconda prompt but I will not supply instructions for those here.

Run Suite2p

• Log into the HPC
• Change directory into code folder

cd basic-uq-hpc-scripts

• Access to all RDP drives on the HPC is through the /QRISdata/ directory. e.g. Q4414 (the pipeline drive) is located at /QRISdata/Q4414/.
• Find the fish that you wish to run Suite2p on in your RDM drive, you can use the ls command to list files in a directory e.g.

ls /QRISdata/Q4414/

• Update the fish_input_path and fish_output_path parameters in the ops file (default is ops_1P_whole.json)
• A full path to a fish might look like

/QRISdata/Q2396/SPIM_120170/Spontaneous/AG_PipelineAuditory_20200910_scn1lab_fish01_spon_2Hz_range245_step5_exposure10_power60_range245_step5_exposure10_power60/

• Update the job name and other HPC parameters in the suite2p_hpc.pbs file (optional)
• Submit the job file to the HPC queue

qsub suite2p_hpc.pbs

• To check the status of jobs you can use qstat
• The HPC will log any errors in files in the HPC home directory, you can review these when things don't work.

Check Suite2p ran correctly

• Log into the HPC
• go into the basic-uq-hpc-scripts folder
• run command:

python write_fish_stats.py <suite2p_output_path>

Where <suite2p_output_path> is the output path for suite2p. It should look something like this: python write_fish_stats.py /QRISdata/Q4414/SPIM/suite2p_220217_JA_fish04_step5/

• Show new text file in the suite2p output folder (the folder with plane0, plane1, ... ) with:

cat <suite2p_path>/s2p_stats.txt

Suite2p didn't work if the "missing files" column in that file has values greater than 0.

Run ANTs

• log onto hpc
• Change directory into code folder

cd basic-uq-hpc-scripts

• Edit variables to set at top of run_ANTs.pbs script
• Note: filenames must have a step size in the filenames or ANTs will fail.
• Launch job

qsub run_ANTs.pbs

• As with suite2p qstat to check the state of a running (or queued) job, output and errors will be written to the HPC home directory

Check ANTs ran correctly

• log onto hpc
• Change directory into code folder

cd basic-uq-hpc-scripts

• Check the .e files using command 'ls'
• Open the error file from a fish that did not run through ANTs correctly using command 'vim'.

Kill a job

• log onto hpc
• Find the job number with command "qstat"
• Kill the job with command "qdel"

qdel JOBNUMBER