trace_combinator

The objective of trace_combinator is to combine traces from different ROIs or different experiments.

• This script takes JSON file with folders where datasets are stored.

• It searches for Trace files with the expected methods, loads them, and

• combines them into a single table that is outputted to the buildPWDmatrix folder.

Invoke

trace_combinator can be run from the command line in two manners:

• Provide input parameters in a folders2Load.json dictionary (see example below) at execution folder or at a rootFolder provided as argument. The script will read ALL the Trace files in the paths provided in folders2Load.json and will combine them into a unique Trace_ files. The output of trace_combinator will be generated at the rootFolder in the directory combined_traces that will contain a new Trace file with the combined traces.

  warning: trace_combinator will combine ALL traces found in the buildPWDmatrix folders within folders2load.json. If you want to filter which Trace files are used, we recommend you to use the second method described below. Alternatively, you can use the --method command-line argument to filter which Trace files are used. In that case, make sure to verify the terminal output to ensure the right Trace files have been used.

• An alternative way to use trace_combinator is to use pipes. In this case, a list of Trace files to process is piped into trace_combinator and the --pipe argument is invoked. Piping can be either done using a command or by sending a list in a file. See examples below.

  Example 1: using ls

  # use ls to select which files you want to combine
  $ ls folder1/Trace*3D*6.ecsv
  folder1/Trace_3D_barcode_KDtree_ROI:6.ecsv  folder1/Trace_3D_barcode_mask:DAPI_ROI:6.ecsv  folder1/Trace_3D_barcode_mask:mask0_ROI:6.ecsv
  
  # then pipe these files into trace_combinator
  $ ls folder1/Trace*3D*6.ecsv | trace_combinator --pipe
  

  this will process the three Trace files listed using ls.

  Example 2: using cat

  # first make list of files to process and write it in a file
  $ cat files_to_combine
  folder1/Trace_3D_barcode_KDtree_ROI:15.ecsv
  folder2/Trace_3D_barcode_KDtree_ROI:6.ecsv
  
  # then pipe these files into trace_combinator
  $ cat files_to_combine | trace_combinator --pipe
  

  this will process the files within files_to_combine.

Use trace_combinator to select traces with specific labels

trace_combinator can also be used to combine traces with specific, user-provided labels, which can be attributed using trace_selector (see above). These can be indicated by using the --label command-line argument. In addition, you need to indicate whether you want to use only the traces that contain the label (labeled), the traces that do not contain the label (unlabeled), or all the traces (all) irrespective of whether they contain or not the label. It is good practice to verify the label column of your output Trace_ file to check that you selected the expected traces.

Inputs

Name shape	Quantity	Mandatory	Description
parameters.json	1	Yes	Parameter file.

Outputs

Name shape	Quantity	Description

Relevant options

usage: trace_combinator [-h] [-F ROOTFOLDER] [-P PARAMETERS] [-A LABEL]
                           [-W ACTION] [--saveMatrix] [--ndims NDIMS]
                           [--method METHOD] [--pipe]

optional arguments:
  -h, --help            show this help message and exit
  -F ROOTFOLDER, --rootFolder ROOTFOLDER
                        Folder with images
  -P PARAMETERS, --parameters PARAMETERS
                        Provide name of parameter files. folders2Load.json
                        assumed as default
  -A LABEL, --label LABEL
                        Add name of label (e.g. doc)
  -W ACTION, --action ACTION
                        Select: [all], [labeled] or [unlabeled] cells plotted
  --saveMatrix          Use to load matlab formatted data
  --ndims NDIMS         Dimensions of trace
  --method METHOD       Method or mask ID used for tracing: KDtree, mask, DAPI
  --pipe                inputs Trace file list from stdin (pipe)

folders2Load.json template

{
    "dataset_name": {
        "Folders": [
            "/user_disk/data/Experiment_4/0_Embryo/buildsPWDmatrix",
            "/user_disk/data/Experiment_4/1_Embryo/buildsPWDmatrix",
            "/user_disk/data/Experiment_4/2_Embryo/buildsPWDmatrix",
            "/user_disk/data/Experiment_4/4_Embryo/buildsPWDmatrix",
            "/user_disk/data/Experiment_4/5_Embryo/buildsPWDmatrix"
        ],
        "PWD_clim": 1.4,
        "PWD_mode": "median",
        "iPWD_clim": 6,
        "iPWD_mode": "median",
        "ContactProbability_scale": 15,
        "ContactProbability_cmin": 0.0,
        "ContactProbability_distanceThreshold": 0.25
    }
}