Db and Settings for Suite2p

Suite2p can be run with different configurations using the db and settings dictionaries. The db dictionary contains recording specific parameters, and the settings dictionary contains pipeline parameters. Here is a summary of all the parameters that the pipeline takes and their default values.

db.npy

Key	GUI Name	Type	Default	Description
data_path	Data path	<class 'list'>	[]	List of folders with tiffs or other files to process.
look_one_level_down	Look one level down	<class 'bool'>	False	Whether to look in all subfolders of all data_path folders when searching for tiffs.
input_format	Input format	<class 'str'>	tif	Can be ['tif', 'h5', 'nwb', 'bruker', 'movie', 'dcimg'].
keep_movie_raw	Keep movie raw	<class 'bool'>	False	Whether to keep binary file of non-registered frames.
nplanes	Number of planes	<class 'int'>	1	Each tiff / file has this many planes in sequence.
nrois	Number of ScanImage ROIs	<class 'int'>	1	Each tiff / file has this many different ROIs.
nchannels	Number of channels	<class 'int'>	1	Specify one- or two- channel recording.
swap_order	Swap the order of channels and planes for multiplexed mesoscope recordings.	<class 'bool'>	False	Swap the order of channels and planes for multiplexed mesoscope recordings.
functional_chan	Functional channel	<class 'int'>	1	This channel is used to extract functional ROIs (1-based).
lines	Lines for each Scanimage ROI	<class 'list'>	None	Line numbers for each ScanImage ROI.
dy	Y position for each Scanimage ROI	<class 'list'>	None	Y position for each ScanImage ROI.
dx	X position for each Scanimage ROI	<class 'list'>	None	X position for each ScanImage ROI.
ignore_flyback	Ignore flyback	<class 'list'>	None	List of planes to not process (0-based).
subfolders	Subfolders	<class 'list'>	None	If len(data_path)==1, subfolders of data_path[0] to use when look_one_level_down is set to True.
file_list	File list	<class 'list'>	None	List of files to process (default is all files in data_path, only supported with one data_path folder).
save_path0	save_path0	<class 'str'>	None	Directory to store results, defaults to data_path[0].
fast_disk	Fast disk	<class 'str'>	None	Directory to store temporary binary file (recommended to be SSD), defaults to save_path0.
save_folder	Save folder	<class 'str'>	suite2p	Directory within save_path0 to save results.
h5py_key	h5py key	<class 'str'>	data	Key in h5py where data array is stored.
nwb_driver	nwb driver	<class 'str'>	``	Driver for nwb file (nothing if file is local).
nwb_series	nwb series	<class 'str'>	``	TwoPhotonSeries name, defaults to first TwoPhotonSeries in nwb file.
force_sktiff	Force tifffile reader	<class 'bool'>	False	Use tifffile for tiff reading instead of scanimage-tiff-reader.
bruker_bidirectional	Bruker bidirectional	<class 'bool'>	False	Tiffs in 0, 1, 2, 2, 1, 0 ... order.
batch_size	Batch size	<class 'int'>	500	Number of frames per batch when writing binary files.

settings.npy

general settings

Key	GUI Name	Type	Default	Description
torch_device	Torch device	<class 'str'>	cuda	Torch device using GPU ('cuda') or CPU ('cpu').
tau	Ca timescale	<class 'float'>	1.0	Timescale for deconvolution and binning in seconds.
fs	Sampling frequency	<class 'float'>	10.0	Sampling rate per plane.
diameter	Diameter	<class 'list'>	[12.0, 12.0]	ROI diameter in Y and X pixels for sourcery and cellpose detection.

run

Key	GUI Name	Type	Default	Description
do_registration	Do registration	<class 'int'>	1	Whether to motion register data (2 forces re-registration).
do_regmetrics	Compute reg metrics	<class 'bool'>	True	Whether or not to compute registration metrics (requires 1500 frames).
do_detection	Do ROI detection	<class 'bool'>	True	Whether or not to run ROI detection and extraction.
do_deconvolution	Do spike deconvolution	<class 'bool'>	True	Whether or not to run spike deconvolution.
multiplane_parallel	Multiplane parallel	<class 'bool'>	False	Whether or not to run each plane as a server job.

io

Key	GUI Name	Type	Default	Description
combined	Combine planes	<class 'bool'>	True	Combine multiple planes after processing into a single result / single canvas for GUI.
save_mat	Save mat	<class 'bool'>	False	Whether to save output as matlab file.
save_NWB	Save NWB	<class 'bool'>	False	Whether to save output as NWB file.
save_ops_orig	Save ops orig	<class 'bool'>	True	Whether to save db, settings, reg_outputs, detection_outputs into ops.npy.
delete_bin	Delete binary	<class 'bool'>	False	Whether to delete binary file after processing.
move_bin	Move binary	<class 'bool'>	False	If True, and fast_disk is different than save_path, binary file is moved to save_path.

registration

Key	GUI Name	Type	Default	Description
align_by_chan2	Align by chan2 (non-func)	<class 'bool'>	False	When two-channel, you can align by non-functional channel (called chan2).
nimg_init	# of frames for refImg	<class 'int'>	400	Number of subsampled frames for finding reference image - choose more if reference image is poor.
maxregshift	Max registration shift	<class 'float'>	0.1	Max allowed registration shift, as a fraction of frame max(width and height).
do_bidiphase	Compute bidiphase offset	<class 'bool'>	False	Whether or not to compute bidirectional phase offset from recording and apply to all frames in recording (applies to 2P recordings only).
bidiphase	Bidiphase offset	<class 'float'>	0.0	Bidirectional phase offset from line scanning (set by user). Applied to all frames in recording.
batch_size	# of frames per batch	<class 'int'>	100	Number of frames per batch - choose fewer if using GPU and running out of memory.
nonrigid	Use nonrigid registration	<class 'bool'>	True	Whether to use nonrigid registration.
maxregshiftNR	Nonrigid max pixel shift	<class 'int'>	5	Maximum pixel shift allowed for nonrigid, relative to rigid, may need to increase value for unstable recordings.
block_size	Nonrigid block size	<class 'tuple'>	(128, 128)	Block size for non-rigid registration (** keep this a multiple of 2, 3, and/or 5 **).
smooth_sigma_time	Time smoothing	<class 'float'>	0	Gaussian smoothing in time to compute registration shifts (may be necessary with low SNR).
smooth_sigma	Smoothing in XY	<class 'float'>	1.15	Gaussian smoothing in XY; ~1 good for 2P recordings, 3-5 may work well for 1P recordings.
spatial_taper	Edge tapering width	<class 'float'>	3.45	Edge tapering width in pixels, may want larger for 1P recordings.
th_badframes	Bad frame threshold	<class 'float'>	1.0	Determines which frames to exclude when determining cropping - set it smaller to exclude more frames, particularly if crop seems small.
norm_frames	Normalize frames	<class 'bool'>	True	Normalize frames when detecting shifts.
snr_thresh	Nonrigid SNR threshold	<class 'float'>	1.2	If any nonrigid block is below this threshold, it gets smoothed until above this threshold. 1.0 results in no smoothing.
subpixel	Nonrigid subpixel reg	<class 'int'>	10	Precision of subpixel registration for nonrigid (1/subpixel steps).
two_step_registration	Run registration twice	<class 'bool'>	False	Whether or not to run registration twice (useful for low SNR data). Set keep_movie_raw to True if setting this parameter to True.
reg_tif	Save registered tiffs	<class 'bool'>	False	Whether to save registered tiffs.
reg_tif_chan2	Save chan2 registered tiffs	<class 'bool'>	False	Whether to save chan2 registered tiffs.

detection

Key	GUI Name	Type	Default	Description
algorithm	Detection algorithm	<class 'str'>	sparsery	Algorithm used for cell detection ['sparsery', 'sourcery', 'cellpose'].
denoise	Denoise	<class 'bool'>	False	Whether to use PCA denoising for cell detection.
block_size	Denoise block size	<class 'tuple'>	(64, 64)	Block size for denoising.
nbins	Max binned frames	<class 'int'>	5000	Max number of binned frames for cell detection (may need to reduce if reduced RAM).
bin_size	Bin size	<class 'int'>	None	Size of bins for cell detection (default is tau * fs).
highpass_time	Highpass time	<class 'int'>	100	Running mean subtraction across bins with a window of size highpass_time (may want to use low values for 1P).
threshold_scaling	Threshold scaling	<class 'float'>	1.0	Adjust the automatically determined threshold in sparsery and sourcery by this scalar multiplier - set it smaller to find more cells.
npix_norm_min	Min npix norm	<class 'float'>	0.0	Minimum npix norm for ROI (npix_norm = per ROI npix normalized by highest variance ROIs' mean npix).
npix_norm_max	Max npix norm	<class 'float'>	100	Maximum npix norm for ROI (npix_norm = per ROI npix normalized by highest variance ROIs' mean npix).
max_overlap	Max overlap	<class 'float'>	0.75	ROIs with more overlap than this fraction with other ROIs are discarded.
soma_crop	Soma crop	<class 'bool'>	True	Crop dendrites from ROI to determine ROI npix_norm and compactness.
chan2_threshold	Chan2 threshold	<class 'float'>	0.25	IoU threshold between anatomical ROI and functional ROI to define as 'redcell'
cellpose_chan2	Cellpose chan2	<class 'bool'>	False	Use Cellpose to detect ROIs in anatomical channel and overlap with functional ROIs
sparsery_settings	N/A	N/A	N/A	N/A
sourcery_settings	N/A	N/A	N/A	N/A
cellpose_settings	N/A	N/A	N/A	N/A

classification

Key	GUI Name	Type	Default	Description
classifier_path	Classifier path	<class 'str'>	None	Path to classifier file for ROIs (default is ~/.suite2p/classifiers/classifier_user.npy).
use_builtin_classifier	Use built-in classifier	<class 'bool'>	False	Use built-in classifier (classifier.npy) instead of user classifier (classifier_user.npy) for ROIs.
preclassify	Pre-classify	<class 'float'>	0.0	Remove ROIs with classifier probability below preclassify before extraction to minimize overlaps

extraction

Key	GUI Name	Type	Default	Description
snr_threshold	SNR threshold	<class 'float'>	0.0	SNR threshold for ROIs.
batch_size	Batch size	<class 'int'>	500	Batch size for extraction.
neuropil_extract	Extract neuropil	<class 'bool'>	True	Whether or not to extract neuropil; if False, Fneu is set to zero.
neuropil_coefficient	Neuropil coefficient	<class 'float'>	0.7	Coefficient for neuropil subtraction.
inner_neuropil_radius	Inner neuropil radius	<class 'int'>	2	Number of pixels to exclude from neuropil next to ROI.
min_neuropil_pixels	Min neuropil pixels	<class 'int'>	350	Minimum number of pixels in the per ROI neuropil.
lam_percentile	Lambda percentile	<class 'float'>	50.0	Percentile of ROI lam weights to ignore when excluding cell pixels for neuropil extraction.
allow_overlap	Allow overlap	<class 'bool'>	False	Pixels that are overlapping are thrown out (False) or used for both ROIs (True).
circular_neuropil	Circular neuropil	<class 'bool'>	False	Force neuropil_masks to be circular instead of square (slow).

dcnv preprocess

Key	GUI Name	Type	Default	Description
baseline	Baseline type	<class 'str'>	maximin	Method for baseline estimation ['maximin', 'prctile', 'constant'].
win_baseline	Baseline window	<class 'float'>	60.0	Window (in seconds) for max filter.
sig_baseline	Baseline sigma	<class 'float'>	10.0	Width of Gaussian filter in frames (applied to find constant or before maximin filter).
prctile_baseline	Baseline percentile	<class 'float'>	8.0	Percentile of trace to use as baseline if using 'prctile' for baseline.