Run suite2p pipeline¶

Overview¶

suite2p is a widely used toolbox for processing 2-photon (2P) calcium imaging datasets.
The IDEAS Run suite2p pipeline tool is a wrapper for the end-to-end pipeline suite2p API function run_s2p().

Run suite2p pipeline comprises the following 7 processing steps:

1. Binary conversion: the input movie(s) is (are) converted into a single suite2p binary file (BinaryRWFile class, namely).

2. Registration (aka motion correction): each frame is aligned to a reference frame via phase correlation, in a rigid and optionally non-rigid way, to correct for motion artifacts.

3. ROI detection: Regions-Of-Interest (ROIs) are detected using computationally efficient, semi-constrained matrix factorization.

4. ROI extraction: neuropil-corrected fluorescence traces and corresponding neuropil fluorescence traces are extracted from the registered movie for all detected ROIs.

5. ROI classification: extracted ROIs, potentially belonging to different cell compartments (cell body aka soma, dendrites, and axon), are curated using either a built-in classifier or a user-provided one, to only keep cell bodies (aka somata) for further analyses.

6. Spike deconvolution: underlying spike events are inferred from the neuropil-corrected fluorescence traces using spike deconvolution based on the OASIS algorithm.

7. Output conversion: the neuropil-corrected fluorescence traces, neuropil fluorescence traces, deconvolved spikes, extraction statistics, classification labels and parameters information can be saved as 4 different formats: .zip containing suite2p's .npy outputs, .isxd cellset and eventset, .nwb and .mat.

Code¶

suite2p documentation¶

For more information and source code from official suite2p sources, follow the links in the table below:

Input files¶

As of version 1.0.0, Run suite2p pipeline supports 3 types of input 2P movies:

import numpy as np
ops = np.load(<path_to_input_ops_npy>, allow_pickle=True).item()
ops[<parameter_name>] = <parameter_value>
np.save(<path_to_output_ops_npy>, ops)

import numpy as np
ops = np.load("ops.npy", allow_pickle=True).item()
ops["high_pass"] = 200
np.save("ops_modified.npy", ops)

Parameters¶

Algorithm overview¶

Run suite2p pipeline processes 2P data through 7 steps illustrated in the flowchart below:

Note that you can also achieve the same workflow by executing the invididual steps separately. This approach offers more modularity and exposure to all parameters in the IDEAS Analysis Tables. In this case, the flow chart looks like:

Output files¶

You can choose the format(s) of Run suite2p pipeline's output among 4 distinct, non-mutually exclusive options:

Each output file has metadata and previews attached to it.
For suite2p .zip, .nwb and .mat outputs, you can explore 8 preview figures and 1 preview movie, illustrating various aspects of the processing pipeline.
You can find below an example of these previews obtained from processing a 5-minute 2P movie of mouse cortex (data courtesy of Dr. Ahmet Arac, MD, at UCLA):

For Inscopix .isxd outputs, you can explore 2 preview figures for cellset_raw.isxd:

and 2 for eventset.isxd (also obtained from processing the same 5-minute 2P movie as above):

Next steps¶

If you choose to output to .isxd, there are many compatible downstream IDEAS tools for further processing and analysis, including combining with other experimental and behavioral data:

• Data quality evaluation and cell filtering

  • Calcium Imaging QC Report
  • Automated Accept/Reject Cells

• Analysis

  • Compare Neural Activity Across States
  • Compare Neural Circuit Correlations Across States
  • Peri-Event Analysis Workflow
  • Detect Events
  • Deconvolve Traces

FAQ¶