11. astec_embryoproperties

astec_embryoproperties can be used to

• either, with a parameter file, extract cell properties from a co-registered image sequence,

• or manipulate a property file (pkl or xml), in order to

  • convert an xml file into a pkl one (or conversely)

  • extract some features from a property file

  • …

11.1. astec_embryoproperties additional options

The following options are available:

-i files ...

input files (pkl or xml) to be read

-o files ...

output files (pkl or xml) to be read

-c files ...

files (pkl or xml) to be compared to those given by -i

-feature features ...

features to be extracted from the input files, that are to be written in the output files. Features have to be chosen in ‘lineage’, ‘h_min’, ‘volume’, ‘surface’, ‘sigma’, ‘label_in_time’, ‘barycenter’, ‘fate’, ‘all-cells’, ‘principal-value’, ‘name’, ‘contact’, ‘history’, ‘principal-vector’, ‘name-score’, ‘cell-compactness’

-property features ...

same as -feature

--diagnosis

performs some test on the read properties

--diagnosis-minimal-volume DIAGNOSIS_MINIMAL_VOLUME

displays all cells with volume smaller than DIAGNOSIS_MINIMAL_VOLUME

--diagnosis-items DIAGNOSIS_ITEMS

minimal number of items to be displayed

-write-selection, --write-selection

convert xml selections into morphonet files

--print-content

print the keys of the input file(s) (read as python dictionary)

--print-keys

same as --print-content

--print-types

print types of read features (for debug purpose)

11.2. Extracting properties from a co-registered image sequence

When a parameter file is passed after the -p option, astec_embryoproperties will compute image sequence properties. Computing cell related informations as well as the lineage tree requires that the (post-corrected) segmentation images have already been co-registered (with astec_intraregistration see section astec_intraregistration). astec_embryoproperties will parse the INTRAREG/INTRAREG_<EXP_INTRAREG>/ directory, and will compute the properties from the images in the POST/POST_<EXP_POST>/ sub-directory, if existing, else of from the SEG/SEG_<EXP_SEG>/ sub-directory.

11.3. Embryo properties output data

The results are stored in the POST/POST_<EXP_POST>/ or SEG/SEG_<EXP_SEG>/ sub-directory under the INTRAREG/INTRAREG_<EXP_INTRAREG> where <EXP_INTRAREG> is the value of the variable EXP_INTRAREG (its default value is 'RELEASE'). The resulting properties will be stored in the same directory than the images they are issued. It will be stored as a pickle python file, and also as a XML file. Both files contain exactly the same information.

According that the POST/POST_<EXP_POST>/ sub-directory exists (that post-corrected segmentation images have been co-registered), 3 files will be created, named after <EN>

/path/to/experiment/
├── ...
├── INTRAREG/
│  └── INTRAREG_<EXP_INTRAREG>/
│     ├── ...
│     ├── POST/
│     │   └── POST_<EXP_POST>/
│     │       ├── <EN>_intrareg_post_lineage.pkl
│     │       ├── <EN>_intrareg_post_lineage.txt
│     │       └── <EN>_intrareg_post_lineage.xml
│     ...
...

The computed information (see section Image sequence properties) are

• cell_barycenter

  Cell center of mass (in voxel coordinates)

• cell_contact_surface

  For each cell, give for each neighboring cell the contact surface. The sum of these contact surfaces is the cell surface.

• cell_principal_vectors

  The cell principal vectors are issued from the diagonalization of the cell covariance matrix (in voxel unit).

• cell_principal_values

  The cell principal value are issued from the diagonalization of the cell covariance matrix (in voxel unit).

• cell_volume (if required)

  Cell volume (in voxel unit)

• cell_surface

  Cell surface (in pixel unit). For this computation, is mandatory that the co-registered images are isotropic (the same voxel size along the 3 dimensions X, Y, and Z).

• cell_lineage (if required)

The text file <EN>_intrareg_post_lineage.txt contains diagnosis information about the sequence. It lists

• the cell with the smallest sizes as well as the ones with the largest sizes

• the cell with a weird lineage: cells without a mother cell, or cells without daughter cells or having more than 2 daughter cells

• cells having a small intersection with its mother cell with respect to either the mother cell volume or the cell volume.

Note that a property file may contain some other information that can be computed either by astec_embryoproperties (e.g. with the --compute-fate option) or by other means.

11.4. Handling property files

astec_embryoproperties can also help managing property files.

• Converting from xml to pkl and the other way around.

  $ astec_embryoproperties -i file.pkl -o file.xml
  

  convert the pickle file file.pkl into the xml file file.xml

• Converting the lineage information from either an xml or an pkl file to a tlp 1 file for lineage visualization

  $ astec_embryoproperties -i file.pkl -o file.tlp
  

  convert the pickle file file.pkl into the tlp file file.tlp

• Merging files.

  $ astec_embryoproperties -i file1.pkl file2.xml ... filen.pkl -o merge.xml merge.pkl
  

  will merge the files file1.pkl, file2.xml , …, filen.pkl (note that they can be either xml or pkl) and write the result both in xml and pkl formats.

• Extracting properties.

  $ astec_embryoproperties -i file.pkl -feature volume surface -o file.xml
  

  will extract the cell volume and surface information from the pickle file file.pkl and write them into the xml file file.xml.

• Comparing property files may help to view changes and/or correction between two property files

  $ astec_embryoproperties -i file.pkl -c file_to_be_compared_to.pkl
  

  compare the two files file.pkl and file_to_be_compared_to.pkl. The comparison is made on all common properties (according it has been implemented). The -feature option allows to select the features to be compared.

• Assessing a property file

  $ astec_embryoproperties -i file.pkl --diagnosis
  

  will run some test/diagnosis on some properties (only a few features are tested). It may help at detecting errors either in the segmented images or in the property file. The -feature option allows to select the features to be tested.

11.5. Diagnosis on property file

Apart reporting diagnosis in the console and in the log file, selection properties (in the morphonet sense) may be added in the output property file (if specified thanks to the -o option) that can be also written as selection files (if the option -write-selection is used).

11.5.1. Contact surfaces

This diagnosis checks whether there are branches with large contact surface distance between consecutive cells (large means larger than maximal_contact_distance, see section Diagnosis parameters). A morphonet selection (of float type) is created whose values are the distance value for both consecutive points that participate in the distance calculation.

11.5.2. Lineage

This diagnosis checks whether the lineage is well-formed.

A morphonet selection (of selection type) is created whose values are

• 10 for the first cell of lineage trees starting after the first time point,

• 20 for cells with multiple mother cells,

• 30 for the last cell of branches ending before the last time point,

• 40 for dividing cells with more than 2 daughter cells,

• 50 for the first cell of short non-terminal branches (short means less than the value of minimal_length, see section Diagnosis parameters).

11.5.3. Name

This diagnosis checks whether the cell names obey to Conklin’s syntax [Con5.].

A morphonet selection (of selection type) is created whose values are

• 10 for cells that are not named but have a non-dividing predecessor that is named

• 20 for cells that are named but have an unamed predecessor

• 30 for cells that are named but differently than their non-dividing predecessor

• 40 for named cells that come after a dividing cell, with a name that does not follow Conklin’s syntax

• 50 for cells that come after a dividing cell, with a name that follows Conklin’s syntax, but with a sibling without name

• 60 for cells that come after a dividing cell, with a name that follows Conklin’s syntax, but with a sibling that has the same name

• 70 for cells that come after a dividing cell, with a name that follows Conklin’s syntax, but with a sibling name that does not follow Conklin’s syntax

• 80 for other errors

11.5.4. Volume

This diagnosis checks whether

• there are small cell volume (small means less than minimal_volume, see section Diagnosis parameters).

• there are branches with a large volume variation (large means larger than maximal_volume_variation, see section Diagnosis parameters). Branch volume variation is computed by \(100 * \frac{\max_{t} v(c_t) - \min_{t} v(c_t)}{\mathrm{med}_t v(c_t)}\) where \(v(c_t)\) is the volume of the cell \(c_t\) and \(\mathrm{med}\) is the median value.

• there are branches with large volume derivatives (large means larger than maximal_volume_derivative, see section Diagnosis parameters). The volume derivative along a branch is calculated as \(100 * \frac{v(c_{t+1}) - v(c_{t})}{v(c_{t})}\) where \(t\) denotes the successive acquisition time points. A morphonet selection (of float type) is created whose values are the absolute value of derivative for both points (\(v(c_{t})\) and \(v(c_{t+1})\)) that participate in the derivative calculation.

1

Tulip is a Data Visualization Software, see tulip.labri.fr