Measurements & Quantification

A lot of measurement implemented in Small Fish comes from the bigfish package (published work) for additional informations refer to the github page of the bigfish package.

Contents

1. Quantification
   a. Levels of quantifications
   b. Field of view level
   c. Cell level
   d. Single molecule level
   e. Co-localization level

2. Measurements
   a. Measurements explanations
   b. How to get the measurement you are looking for

1. Quantification

In this section is explained how quantification is performed and in particular the different behaviours of the software when given cell segmentation or not.

Levels of quantification

In Small Fish quantification is organised on different levels :

• Field of view (fov level)
• Cellular level (cell level)
• Single molecule level (spot level)
• Co-localization level

Each of this level can results in one (or two for co-localization) data file containing a different scale of information. As an example the field of view data file will contain one line of measurements for every field of view that were processed whereas the cell data file will contain one line of measurements for every cell that were computed. It is then possible to link cells to their belonging field of view using unique identifiers, in this case named 'acquisition_id'.

Let's say we have computed 20 cells in our field of. In the fov datafile they will be 1 line reading acquisition_id = 1 but in the cell data file we will find 20 lines all reading acquisition_id = 1 but with a different cell_id.

Field of view level

The field of view level is mandatory it contains a variety of measurements that were taken on the fov as a whole such as total spot number, mean intensity as well as all the parameters given to the software for quantification. At this level one line of measurement represents one image stack the quantification is saved in filename with unique identifier acquisition_id.

It is important to understand that all measurements present in this level are never impacted by the segmentation given to the software and thus can be performed without segmentation at all. That is to say that the total spot number will give you the number of spot detected in the field of view independently of the fact they are found within segmentation masks or not.

Field of view level can generally be used for quantification of rna abundency, single molecule signal intensity, signal to noise ratio or to check what parameters were used.

Cell level

The cell level is optional and requires cell segmentation to be performed it contains measurements on cells as individuals such as cell size, single molecule number inside nucleus, single molecule number inside foci... At this level one line of measurement represents one cell and the quantification is saved in filename_cell_result.

If you are not familiar with segmentation labels they consist in re-asignining an integer value to each pixel. Pixels with value 0 are considered background and pixels forming a cell are assigned a value superior to 0. By assigning a different pixel values to different cells we are able to distinguish one cell from another, this integer value is usually call label but it Small Fish it is called cell_id. Therefore it is important to remember that cell id is only unique within the same field of view meaning that the couple (acquisition_id, cell_id) is unique. This behavior is intended to allow matching cells during co-localization measurement.

Cell level quantification is the biggest part of Small Fish and and offer a large variety of measurements allowing user to quantify :

• nuclear single molecule abundancy
• cytoplasm single molecule abundancy
• spatial distribution of cytoplasmic spot respect to nuclei, membrane, protrusion...
• cell and nuclei size
• signal to noise ratio
• statics on signal in nuclei/cell
• single molecule polarization (PI)
• single molecule dispertion (DI)
• single molecule peripheral distribution (PDI)
• foci quantification

For cells the quantification is restrained to segmentation masks meaning that all spots, all clusters found outside of masks will be ignored and not refleted in cell measurements. Cell processing is handled by bigfish package it consists in extracting information from each cells individually using label segmentation label to that end cell are individually cropped from main image along with their nucleus and cytoplasm mask allowing ligther data processing during measurements computation. Used cropping is saved for each cell with name 'cell_bbox'.

Single molecule level

The spot level is optional and does not require cell segmentation it is computed when the user enables the spot extraction option during detection and gives a valid path for saving. It considered single molecules as individuals and measure for each of them their intensity as well to which cluster it belongs, to which cell it belongs and if it is a nuclear spot. Contrary to other data files the spot data file is saved during quantification as it allows the software to not keep in memory information on intensities of each spots of each field of view computed.

This data file can become much more heavy since it contains one line per spot detected, it is saved as filename_spot_extraction with unique identifier spot_id.

Co-localization level

Co-localization with Small Fish allows the user to quantifiy spatial proximity between two population of spots detected in seperate acquisitions. As an example if I want to quantify my multichannel image stack containing two different fluroescent reporters in two distinct channels I can start by segmenting cells followed with two round of detection (1 per channel) resulting in two acquisitions. Co-localization measurement will then enable me to quantify, cell by cell, the number, or fraction of single molecule from channel 1 co-localizing with channel 2.

Any time co-localization is tested the measurement of co-localization is taken both ways since testing the co-localization of a population of single molecule A with another population B is different than testing the population B with A (see illustration bellow).

To perform co-localization you can either perform at least 2 detections or load spot extraction. In the case when both selected acquisitions have cell level quantification (i.e cells were segmented) a cell to cell co-localization is performed otherwise a global co-localization is performed.

Global co-localization

Global co-localization is achieved when at least one tested acquisition was not computed on cell level this will restrain the quantification to yield only one co-colocalization measurement per couple of acquisition tested. While this can still give you hints or ideas about the behaviour of your system a cell to cell quantification will be statiscally relevant as it will give a better signification to your mean and standard deviation values.

When saving results an additional data file will be created namely global_coloc_result containing measurements, acquisitions names and ids as well as the colocalisation distance used (see Co-localization section).

Cell to cell co-localization

Cell to cell co-localization comes with a heavier data density which makes it a bit less comfortable to handle. First let's explain how cells are matched in the case of computing cell to cell co-localization with acquisition1 and acquisition2. Cells from acquisition1 and acquisition2 are matched using their cell_id which corresponds to the value of the segmentation label in the cell. Therefore it is important to understand the software will not check if you are performing the co-localization test between two acquisitions segmented identically. Needless to say if you perform co-localization measurement in different field of views or with different segmentation the measurements will be falsed.

When cell to cell co-localization is computed an additional data file named cell2cell_coloc_result is created containing one line of measurements for each cell. **These cells can be linked to cell data file using cell_id and the acquisition_id to link to cell level quantification. This data frame contains also the the colocalisation distance used (see Co-localization section).

Note : Data might be more comfortable to handle with one file per field of fiew in this case. To achieve this save results after testing all co-localization of interest for your fov and then click the reset button before moving on to next fov.

Unique identifiers

In this section you can find a graphical representation of the Small Fish database summing up identifiers across all levels.

2. Measurements

1. Measurements explanation

Signal to noise ration (snr)

Signal to noise measurement is performed using the metric from the fish-quant environnement. It aims at assessing the quality of specific signal confronted to background and is computed using the following formula :

$SNR = \frac{max(spot signal) - mean(background)}{std(background)}$

Cell background

For snr computation and general quantification (i.e quantification without cell segmentation), is considered as cell background signal, signal which is find in a region twice as large as the spot radius around detected spots without including signal from spot (understand signal in spot radius).

index distance

Measures including index distance such as index distance centrosome are normalised with the distance that would be expected if rna were uniformly distributed. In the case of centrosomal distance it means that the mean centrosomal distance is computed for the detected distribution and then divided by the mean of distance of each pixels in the cell.

index polarization, dispertion and peripheral distribution

Those 3 indexes are presented in Stueland M, Wang T, Park HY, Mili S. RDI Calculator: An Analysis Tool to Assess RNA Distributions in Cells. Sci Rep. 2019 Jun 4;9(1):8267. doi: 10.1038/s41598-019-44783-2. PMID: 31164708; PMCID: PMC6547641.
I chose to directly quote the paper for this description.

"The Polarization Index (PI) is calculated by identifying the centroid of the RNA signal and measuring its displacement from the centroid of the cell. This displacement is divided by the radius of gyration, calculated as the root-mean-square distance of all pixels to the centroid of the cell, in order to normalize the polarization to the size and elongation of the cell."

$PI = \frac{\sqrt{(x_{rna}-x_{cell})^2 + (y_{rna}-y_{cell})^2}}{Rg_{cell}}$

"where $x_{rna}$, $y_{rna}$ are the coordinates of the RNA centroid and $x_{cell}$, $y_{cell}$, are the coordinates of the centroid of the cell in the two dimensional image, and $Rg_{cell}$ is the radius of gyration. PI values increase with increased polarization of the RNA signal and with increased distance from the cell centroid."

"To derive the Dispersion Index (DI), the second moment of RNA pixel intensity positions relative to the centroid of the total RNA signal is calculated."

$\mu_2 = \sum_{ij}r^2_{ij} \frac{I_{ij}}{\sum_{ij}I_{ij}} $

"where $r_{ij}$ is the distance of the pixel (i,j) to the centroid of the RNA signal and $I_{ij}$ is the intensity value of pixel (i,j) in the two dimensional image. To normalize for differences in cell morphology, the second moment of the RNA is divided by the second moment of a hypothetical uniform distribution, which is derived as the second moment of all pixels within the binary cell mask image. A completely diffuse RNA has a DI value of 1. The DI value of an RNA that is concentrated in any region within the cell is less than 1 and is inversely correlated with the degree of RNA concentration. An RNA that is distributed towards the cell periphery exhibits a DI value larger than 1, but this is affected by the degree of polarization."

The Peripheral Distribution Index (PDI) is calculated similar to the dispersion index, but in this case the second moment of RNA pixel intensity positions is calculated relative to the centroid of the nucleus14. This metric is not affected by the polarization of the RNA distribution. PDI value is 1 for a completely diffuse RNA, it is less than 1 for a perinuclear RNA and more than 1 for a peripherally distributed RNA

2. How to get the measurement you are looking for

In this section you will find a summup of all measurements availables. Measurement in bold are detailed in section above.

Field of View measurements

See quantification section

Measure	Explanation	Requires segmentation	Requires clustering	Type
acquisition_id	Unique identifier for field of views	No	No	Identifier
name	User given label for field of view	No	No	Parameter
spot_number	Total number of spot detected	No	No	Measure
spotsSignal_median	Median signal of single molecule	No	No	Measure
spotsSignal_mean	Mean signal of single molecule	No	No	Measure
spotsSignal_std	Standard deviation of single molecule signal	No	No	Measure
median_pixel	Median signal in fov	No	No	Measure
mean_pixel	Mean pixel in fov	No	No	Measure
snr_median	Median signal to noise ratio	No	No	Measure
snr_mean	Mean signal to noise	No	No	Measure
snr_std	Standard deviation of signal to noise	No	No	Measure
cell_medianbackground_std	Median value of signal in cells away from spots	No	No	Measure
cell_meanbackground_mean	Mean value of signal in cells away from spots	No	No	Measure
cell_meanbackground_std		No	No	Measure
cell_stdbackground_mean		No	No	Measure
cell_stdbackground_std		No	No	Measure
cluster_number	Number of cluster detected	No	Yes	Measure
total_spots_in_clusters	Number of spots detected in clusters	No	No	Measure
cell_number	Number of cell detected	Yes	No	Measure
segmentation_done	Wheter segmentation was performed or not	No	No	Parameter
cyto_model_name	Name of cellpose model used for cytoplasm segmentation	Yes	No	Parameter
cytoplasm channel	Channel used for cytoplasm segmentation	Yes	No	Parameter
cytoplasm diameter	Object size in pixel given to cellpose for cytoplasm segmentation	Yes	No	Parameter
nucleus_model_name	Name of cellpose model used for nuclei segmentation	Yes	No	Parameter
nucleus channel	Channel used for nuclei segmentation	Yes	No	Parameter
other_nucleus_image	filename of nucleus image in case user is using a second image to perform segmentation	Yes	No	Parameter
nucleus diameter	Object size in pixel given to cellpose for cytoplasm segmentation	Yes	No	Parameter
Segment only nuclei	True/False parameter enabling user to ignore cytoplasm segmentation	Yes	No	Parameter
show segmentation	True/False parameter enabling user to use napari to visualize and edit segmentation	Yes	No	Parameter
saving path filename	full path to saved fodler where segmentation visuals were saved.	Yes	No	Parameter
image path	Full path to loaded image	No	No	Parameter
is_3D_stack	True if loaded image was considered a 3D stack	No	No	Parameter
is_multichannel	True if loaded image was considered a multi-channel stack	No	No	Parameter
do_dense_regions_deconvolution	True if user tried to perform dense region deconvolutions	No	No	Parameter
do_cluster_computation	True if user tried to perform cluster computation	No	No	Parameter
show_napari_corrector	True/False parameter enabling user to use napari to visualize and edit detection	No	No	Parameter
shape	Loaded image shape	No	No	Parameter
dim	Loaded image dimension	No	No	Parameter
reordered_shape	Shape of image after axis re-ordering	No	No	Parameter
threshold	Threshold value used for single molecule detection	No	No	Parameter
threshold penalty	Multiplicative factor applied to automatic threshold	No	No	Parameter
channel_to_compute	Channel on which was performed detection	No	No	Parameter
voxel_size_z	Size of a pixel in nanometer	No	No	Parameter
voxel_size_y	Size of a pixel in nanometer	No	No	Parameter
voxel_size_x	Size of a pixel in nanometer	No	No	Parameter
spot_size_z	Expected single molecule size	No	No	Parameter
spot_size_y	Expected single molecule size	No	No	Parameter
spot_size_x	Expected single molecule size	No	No	Parameter
log_kernel_size_z	Size of Laplacian of Gaussian filter in pixel, optional, will be computed from spot size if not given	No	No	Parameter
log_kernel_size_y	Size of Laplacian of Gaussian filter in pixel, optional, will be computed from spot size if not given	No	No	Parameter
log_kernel_size_x	Size of Laplacian of Gaussian filter in pixel, optional, will be computed from spot size if not given	No	No	Parameter
minimum_distance_z	Minimum distance between detected spots, optional, will be computed from spot size if not given	No	No	Parameter
minimum_distance_y	Minimum distance between detected spots, optional, will be computed from spot size if not given	No	No	Parameter
minimum_distance_x	Minimum distance between detected spots, optional, will be computed from spot size if not given	No	No	Parameter
alpha	Dense region deconvolution parameter, distribution percentile for reference spot building	No	No	Parameter
beta	Dense region deconvolution parameter, multiplicative factor applied to distribution median for bright pixel thresholding	No	No	Parameter
gamma	Dense region deconvolution parameter	No	No	Parameter
deconvolution_kernel_z	Dense region deconvolution parameter, optional, used to specify kernel size (nanometers) for gaussian blur pre-processing	No	No	Parameter
deconvolution_kernel_y	Dense region deconvolution parameter, optional, used to specify kernel size (nanometers) for gaussian blur pre-processing	No	No	Parameter
deconvolution_kernel_x	Dense region deconvolution parameter, optional, used to specify kernel size (nanometers) for gaussian blur pre-processing	No	No	Parameter
cluster size	Expected size of clusters in nanometer, parameter used by DBSCAN algorithm to find core points	No	No	Parameter
min number of spots	Minimum number of spots for single molecules to form a cluster, parameter used by DBSCAN algorithm	No	No	Parameter
nucleus channel signal	Channel given for nucleus releated feature	Yes	No	Parameter
show_interactive_threshold_selector	True/False parameter enabling interactive napari visualization of threshold during detection	No	No	Parameter
spots_extraction_folder	if fullpath is provided spots quantification will be extracted to this location	No	No	Parameter
spots_filename	if provided spots quantification result filename	No	No	Parameter
do_spots_csv	True/False parameter indicating if spots quantification was saved to csv format	No	No	Parameter
do_spots_excel	True/False parameter indicating if spots quantification was saved to csv excel	No	No	Parameter
do_spots_feather	Will be removed in future version	No	No	Parameter
voxel_size	Size of a pixel in nanomter	No	No	Parameter
spot_size	Expected single molecule size	No	No	Parameter
log_kernel_size	Size of Laplacian of Gaussian filter in pixel, optional, will be computed from spot size if not given	No	No	Parameter
minimum_distance	Minimum distance between detected spots, optional, will be computed from spot size if not given	No	No	Parameter
deconvolution_kernel	Dense region deconvolution parameter, optional, used to specify kernel size (nanometers) for gaussian blur pre-processing	No	No	Parameter

Cell measurements

See quantification section

Measure	Explanation	Requires segmentation	Requires clustering	Type
acquisition_id	Identifier for reference to field of view quantification	Yes	No	Identifier
name	User given label for field of view	Yes	No	Parameter
cell_id	Cell identifier not unique amongst different filed of views. Corresponds to the value of this cell's pixel in segmentation label	Yes	No	Identifier
cell_bbox	coordinates of cell bounding boxe	Yes	No	Measure
index_mean_distance_cell	Normalised mean value of distance between detected spots and cell membrane	Yes	No	Measure
index_median_distance_cell	Normalised median value of distance between detected spots and cell membrane	Yes	No	Measure
index_mean_distance_nuc	Normalised mean value of distance between detected spots and nucleus	Yes	No	Measure
index_median_distance_nuc	Normalised median value of distance between detected spots and nucleus	Yes	No	Measure
proportion_rna_in_nuc	Fraction of detected spots found within nucleus	Yes	No	Measure
nb_rna_out_nuc	Number of spots found outside nucleus mask	Yes	No	Measure
nb_rna_in_nuc	Number of spots found inside nucleus mask	Yes	No	Measure
index_polarization		Yes	No	Measure
index_dispersion		Yes	No	Measure
index_peripheral_distribution		Yes	No	Measure
index_rna_nuc_edge	Normalised number of spots detected closter than 500 nanometers from nucleus	Yes	No	Measure
index_rna_nuc_radius_500_1000	Normalised number of spots detected in region spaning from 500 nanometers to 1000 nanometers from nucleus	Yes	No	Measure
index_rna_nuc_radius_1000_1500	Normalised number of spots detected in region spaning 1000 nanometers to 1500 nanometers from nucleus	Yes	No	Measure
index_rna_nuc_radius_1500_2000	Normalised number of spots detected in region spaning 1500 nanometers to 2000 nanometers from nucleus	Yes	No	Measure
index_rna_nuc_radius_2000_2500	Normalised number of spots detected in region spaning 2000 nanometers to 2500 nanometers from nucleus	Yes	No	Measure
index_rna_nuc_radius_2500_3000	Normalised number of spots detected in region spaning 2500 nanometers to 3000 nanometers from nucleus	Yes	No	Measure
index_rna_cell_radius_0_500	Normalised number of spots detected closter than 500 nanometers from cell membrane	Yes	No	Measure
index_rna_cell_radius_500_1000	Normalised number of spots detected in region spaning from 500 to 1000 nanometers from cell membrane	Yes	No	Measure
index_rna_cell_radius_1000_1500	Normalised number of spots detected in region spaning from 1000 to 1500 nanometers from cell membrane	Yes	No	Measure
index_rna_cell_radius_1500_2000	Normalised number of spots detected in region spaning from 1500 to 2000 nanometers from cell membrane	Yes	No	Measure
index_rna_cell_radius_2000_2500	Normalised number of spots detected in region spaning from 2000 to 2500 nanometers from cell membrane	Yes	No	Measure
index_rna_cell_radius_2500_3000	Normalised number of spots detected in region spaning from 2500 to 3000 nanometers from cell membrane	Yes	No	Measure
proportion_rna_nuc_edge	Proportion of spots detected closer than 500 nanometers from nucleus	Yes	No	Measure
proportion_rna_nuc_radius_500_1000	Proportion of spot detected in region spaning from 500 nanometers to 1000 nanometers from nucleus	Yes	No	Measure
proportion_rna_nuc_radius_1000_1500	Proportion of spot detected in region spaning from 1000 nanometers to 1500 nanometers from nucleus	Yes	No	Measure
proportion_rna_nuc_radius_1500_2000	Proportion of spot detected in region spaning from 1500 nanometers to 2000 nanometers from nucleus	Yes	No	Measure
proportion_rna_nuc_radius_2000_2500	Proportion of spot detected in region spaning from 2000 nanometers to 2500 nanometers from nucleus	Yes	No	Measure
proportion_rna_nuc_radius_2500_3000	Proportion of spot detected in region spaning from 2500 nanometers to 3000 nanometers from nucleus	Yes	No	Measure
proportion_rna_cell_radius_0_500	Proportion of spots detected closer than 500 nanometers from cell membrane	Yes	No	Measure
proportion_rna_cell_radius_500_1000	Proportion of spot detected in region spaning from 500 nanometers to 1000 nanometers from cell membrane	Yes	No	Measure
proportion_rna_cell_radius_1000_1500	Proportion of spot detected in region spaning from 1000 nanometers to 1500 nanometers from cell membrane	Yes	No	Measure
proportion_rna_cell_radius_1500_2000	Proportion of spot detected in region spaning from 1500 nanometers to 2000 nanometers from cell membrane	Yes	No	Measure
proportion_rna_cell_radius_2000_2500	Proportion of spot detected in region spaning from 2000 nanometers to 2500 nanometers from cell membrane	Yes	No	Measure
proportion_rna_cell_radius_2500_3000	Proportion of spot detected in region spaning from 2500 nanometers to 3000 nanometers from cell membrane	Yes	No	Measure
proportion_rna_in_foci	Fraction of spots deteted in foci	Yes	No	Measure
proportion_nuc_area	Ratio of nucleus area on total cell area	Yes	No	Measure
cell_area	cell area in squared pixel	Yes	No	Measure
nuc_area	nucleus area in squared pixel	Yes	No	Measure
cell_area_out_nuc	cytoplasm area without nucleus in squared pixel	Yes	No	Measure
nucleus_mean_signal	mean signal in nucleus (from nucleus channel signal)	Yes	No	Measure
nucleus_median_signal	median signal in nucleus (from nucleus channel signal)	Yes	No	Measure
nucleus_max_signal	maximum signal in nucleus (from nucleus channel signal)	Yes	No	Measure
nucleus_min_signal	minimum signal in nucleus (from nucleus channel signal)	Yes	No	Measure
snr_mean		Yes	No	Measure
snr_median		Yes	No	Measure
snr_std		Yes	No	Measure
cell_center_coord	Pixel coordinates of cell centroid	Yes	No	Measure
foci_number	Number of foci detected in cell	Yes	Yes	Measure
foci_in_nuc_number	Number of foci detected in nucleus	Yes	Yes	Measure
clustered_spot_number	Number of spot detected in clusters	Yes	Yes	Measure
free_spot_number	Number of spot detected outside of clusters	Yes	Yes	Measure
total_rna_number	Number of spot detected in cell	Yes	Yes	Measure

Spots measurements

See quantification section

Measure	Explanation	Requires segmentation	Requires clustering	Type
acquisition_id	Identifier refering to field of view quantification	No	No	Identifier
spot_id	Unique indentifier for spots	No	No	Identifier
intensity	Intensity of detected spot	No	No	Measure
cell_label	Label of cell at spot location, this matches cell_id identifier from cell results table	Yes	No	Identifier
in_nucleus	Boolean value indicating if spot was detected inside nucleus	Yes	No	Measure
coordinates	Pixel coordinates of detected spot	No	No	Measure
cluster_id	Index indicating to which cluster the spots belongs. If -1 then spot is not in a cluster	No	Yes	Measure

Global co-localization measurements

See quantification section

Measure	Explanation	Requires segmentation	Requires clustering	Type
name1	User given label to first acquisition of the co-localization pair	No	No	Parameter
name2	User given label to second acquisition of the co-localization pair	No	No	Parameter
acquisition_id_1	Identifier in field of view quantification of first acquisition from the tested pair	No	No	Indentifier
acquisition_id_2	Identifier in field of view quantification of second acquisition from the tested pair	No	No	Identifier
colocalisation_distance	Maximum distance that can separate co-localizing spots	No	No	Parameter
spot1_total	Number of detected spots in first acquisition	No	No	Measure
spot2_total	Number of detected spots in second acquisition	No	No	Measure
fraction_spots1_coloc_spots2	Fraction of spots from first acquisition co-localizing with at least one spot from second acquisition	No	No	Measure
fraction_spots2_coloc_spots1	Fraction of spots from second acquisition co-localizing with at least one spot from first acquisition	No	No	Measure
fraction_spots2_coloc_cluster1	Fraction of spots from second acquisition co-localizing with at least one clustered spot from first acquisition	No	Yes	Measure
fraction_spots1_coloc_cluster2	Fraction of spots from first acquisition co-localizing with at least one clustered spot from second acquisition	No	Yes	Measure
fraction_cluster1_coloc_cluster2	Fraction of clustered spots from first acquisition co-localizing with at least one clustered spot from second acquisition	No	Yes	Measure
fraction_cluster2_coloc_cluster1	Fraction of clustered spots from second acquisition co-localizing with at least one clustered spot from first acquisition	No	Yes	Measure
fraction_spots1_coloc_free2	Fraction of spots from first acquisition co-localizing with at least one free spot from second acquisition	No	Yes	Measure
fraction_spots2_coloc_free1	Fraction of spots from second acquisition co-localizing with at least one free spot from first acquisition	No	Yes	Measure

Cell to cell co-localization measurements

See quantification section

Measure	Explanation	Requires segmentation	Requires clustering	Type
cell_id	Appear as index of frame (rows instead of columns). Value of cell label	No	No	Index
clustered_spot_number		No	No	Measure
clustered_spots_with_clustered_spots_count		No	No	Measure
clustered_spots_with_clustered_spots_fraction		No	No	Measure
spots_with_clustered_spots_count		No	No	Measure
spots_with_clustered_spots_fraction		No	No	Measure
spots_with_spots_count		No	No	Measure
spots_with_spots_fraction		No	No	Measure
total_rna_number		No	No	Measure
voxel_size		No	No	Measure
pair_name		No	No	Measure
acquisition_id_1		No	No	Measure
acquisition_id_2		No	No	Measure
colocalisation_distance		No	No	Measure