Tutorial 5 - Integrating with Suite2p (running Suite2p pipeline and storing results)¶

none

[1]:

ipython3

import imagingplus as ip
from imagingplus.processing import suite2p

import numpy as np

Warning: cellpose did not import
No module named 'cellpose'
cannot use anatomical mode, but otherwise suite2p will run normally

In this tutorial, we add Suite2p processing to an existing expobj using the Suite2p module of Imaging+. We also show how to run the suite2p pipeline directly from within the .Suite2p submodule of the expobj.

Note: it is also possible to directly import independently run Suite2p results for this experiment (this is demonstrated in the final section).

Suite2p provides an excellent cell (ROI) detection algorithm for calcium imaging datasets. The pipeline includes registration and spike deconvolution as well. Please familiarize yourself with the basic functionality of Suite2p using their documentation (found here) before proceeding with these examples below.

Adding the Suite2p module¶

Import data analysis object¶

none

[ ]:

ipython3

# specify the .pkl path to import expobj
prep = 'PS14'
date = '2021-01-28'

expobj: ip.Experiment = ip.import_obj(pkl_path=f'/home/pshah/mnt/qnap/Analysis/{date}/{prep}/{prep}_analysis.pkl')

# set up new Suite2p submodule
expobj.Suite2p = suite2p.Suite2pExperiment(trialsTiffsSuite2p=expobj.Suite2p.tiff_paths_to_use_s2p,
                                          s2pResultsPath=expobj.Suite2p.s2pResultsPath)

Set any necessary bad_frames to Suite2p. These frames will be ignored during the ROI detection step. In this example, we set all frames from trials from the post-4ap experiment group (expGroup) as bad frames. In other experiments, like for example all-optical experiments, we use this feature to set all photostimulation frames as bad_frames

none

[ ]:

ipython3

# add Suite2p bad frames from post4ap trials
expobj.Suite2p.bad_frames = []
for trial in expobj.trialIDs:
    if 'post' in expobj.TrialsInformation[trial]['expGroup']:
        trialobj: ip.TwoPhotonImagingTrial = expobj.load_trial(trialID=trial)

        expobj.Suite2p.add_bad_frames(frames=np.arange(trialobj.Suite2p.trial_frames[0], trialobj.Suite2p.trial_frames[1]), bad_frames_npy_loc = expobj.dataPath)

Prepare for Suite2p run¶

We next need to setup the Suite2p settings to run for our particular imaging experiment. Use the .Suite2p.update_ops(dict) method to update any ops options before running to Suite2p.

To get the necessary imaging metadata parameters, here we use one of the trialobj contained within the present expobj to retrieve that information.

none

[ ]:

ipython3

trialobj: ip.TwoPhotonImagingTrial = expobj.load_trial(trialID=expobj.trialIDs[0])

pix_sz_x = trialobj.imparams.pix_sz_x
pix_sz_y = trialobj.imparams.pix_sz_y
diameter_x = 13 / pix_sz_x
diameter_y = 13 / pix_sz_y
diameter = (int(diameter_x), int(diameter_y)) if int(diameter_y) != int(diameter_x) else int(diameter_x)

Set the new values for Suite2p by using .Suite2p.update_ops(dict)

none

[ ]:

ipython3

# new values for ops dictionary
new_ops = {
    'fs': trialobj.imparams.fps,
    'tau': 0.70,
    'num_workers': 50,
    'diameter': diameter,
    'delete_bin': False,  # temporary not deleting binaries in case needed for further testing!
    'batch_size': 2000,  # larger frames will be more RAM intensive, scale user batch size based on num pixels in 512x512 images
    'save_folder': expobj.suite2p_save_path
}

expobj.Suite2p.update_ops(new_ops)
expobj.save()

print('ops:\n', expobj.Suite2p.ops)
print('db:\n', expobj.Suite2p.db)

Run Suite2p¶

Now, we are all set to run Suite2p. We have provided the .Suite2p.s2pRun method as a wrapper to run the Suite2p on the current expobj.

none

[ ]:

ipython3

# run the suite2p pipeline
expobj.Suite2p.s2pRun(expobj=expobj)

Import suite2p results directly¶

If we had run suite2p on our experiment independent of creating the expobj, we can add that Suite2p output to the present expobj by providing a valid path to the s2pResultsPath argument when setting up the .Suite2p submodule on our expobj.

Note: the output that is required in s2pResultsPath is the path to <../plane0/> of the Suite2p results output.

none

[ ]:

ipython3

# specify the .pkl path to import expobj
prep = 'PS14'
date = '2021-01-28'

expobj: ip.Experiment = ip.import_obj(pkl_path=f'/home/pshah/mnt/qnap/Analysis/{date}/{prep}/{prep}_analysis.pkl')

# set up new Suite2p submodule
expobj.Suite2p = suite2p.Suite2pExperiment(trialsTiffsSuite2p=expobj.Suite2p.tiff_paths_to_use_s2p,
                                          s2pResultsPath='<path/to/suite2p/output>')