After having installed TopoStats you are ready to run it. For convenience TopoStats provides a command line interface topostats that will load a default configuration file and process all images with reasonable default configuration options.

However, because the location of your image files can not be known in advance you must make a copy of the default configuration and modify it to work with your files. This guide will hopefully take you through the process of running TopoStats and customising the configuration file with which it is run. If you encounter any problems please ask questions in the Discussions. If you think you have encountered a bug or have a feature suggestion please create an Issue.

Organising Scans#

You should place all files you wish to batch process in a single directory. They can be nested in separate folders as TopoStats will scan for all images within this directory but currently it will only process one scan type at a time (i.e. .spm or .jpk or .asd). This may change in the future.

Command Line Navigation#

TopoStats currently runs as a command-line programme. To use it you will have to use a “prompt” or “terminal” (they’re essentially the same thing). What you use will depend on your operating system, but the following are some simple commands on navigation. If you use Windows then for consistency it is recommended to install and use PowerShell.

At the command line you use cd to change directory to the location of your files. For example if your scans are on the C-drive in C:\User\me\work\spm\2022-12-08\scans then you would

cd c:/User/me/work/spm/2022-12-08/scans

If you are on a Linux or OSX system then paths are not prefixed with letters and your files may be saved to /home/me/work/spm/2022-12-08/scans. To change directory there you would…

cd /home/me/work/spm/2022-12-08/scans

NB - Always use a forward-slash (/) when typing directory paths. Windows will display back-slash (\) but understands forward-slash. Under Linux and OSX they mean different things and so you should always use forward-slash (/).

You can always find out what location you are at in the command line using the pwd command (print working directory) and it will print out the directory you are currently at.


To navigate up one directory level use cd ... These can be chained together and directories separated with /.

# Move up a single directory level
cd ..
# Move up another two directory levels
cd ../../

You can list the files in a directory using the ls command.


To learn more about the command line see the Introduction to the Command Line for Genomics.

Running TopoStats#

The default location that TopoStats looks for scans is the directory from which it is invoked. Once you start your shell/terminal you will therefore need to do two things.

  1. Navigate to the location of the scans you wish to process using cd /path/to/where/scans/are/located.

  2. Activate the virtual environment under which you installed TopoStats (refer to installed if unsure).

You can now run topostats by invoking topostats process and you should start to see some output similar to that below.

cd /path/to/where/scans/are/located
topostats process
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Configuration is valid.
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Plotting configuration is valid.
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Configuration file loaded from      : None
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Scanning for images in              : /home/neil/work/projects/topostats/TopoStats
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Output directory                    : output
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Looking for images with extension   : .spm
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Images with extension .spm in /home/neil/work/projects/topostats/TopoStats : 32
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Thresholding method (Filtering)     : std_dev
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Thresholding method (Grains)        : std_dev

On a successful completion you should see a message similar to the following which indicates various aspects of the run along with information about how to give feedback, report bugs and cite the software.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ COMPLETE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

  Base Directory              : /home/neil/work/projects/topostats/TopoStats
  File Extension              : .spm
  Files Found                 : 1
  Successfully Processed      : 1 (100.0%)
  Configuration               : output/config.yaml
  All statistics              : output/all_statistics.csv
  Distribution Plots          : output/summary_distributions

  Email                       :
  Documentation               :
  Source Code                 :
  Bug Reports/Feature Request :
  Citation File Format        :

  If you encounter bugs/issues or have feature requests please report them at the above URL
  or email us.

  If you have found TopoStats useful please consider citing it. A Citation File Format is
  linked above and available from the Source Code page.

The command topostats process has a number of additional flags for passing different options. These can be viewed using the -h or --help flag.

  topostats process --help
usage: topostats process [-h] [-c CONFIG_FILE] [-s SUMMARY_CONFIG] [--matplotlibrc MATPLOTLIBRC] [-b BASE_DIR] [-j CORES] [-l LOG_LEVEL] [-f FILE_EXT] [--channel CHANNEL] [-o OUTPUT_DIR] [--save-plots SAVE_PLOTS] [-m MASK] [-w WARNINGS]

Process AFM images. Additional arguments over-ride those in the configuration file.

  -h, --help            show this help message and exit
  -c CONFIG_FILE, --config-file CONFIG_FILE
                        Path to a YAML configuration file.
  -s SUMMARY_CONFIG, --summary-config SUMMARY_CONFIG
                        Path to a YAML configuration file for summary plots and statistics.
  --matplotlibrc MATPLOTLIBRC
                        Path to a matplotlibrc file.
  -b BASE_DIR, --base-dir BASE_DIR
                        Base directory to scan for images.
  -j CORES, --cores CORES
                        Number of CPU cores to use when processing.
  -l LOG_LEVEL, --log-level LOG_LEVEL
                        Logging level to use, default is 'info' for verbose output use 'debug'.
  -f FILE_EXT, --file-ext FILE_EXT
                        File extension to scan for.
  --channel CHANNEL     Channel to extract.
  -o OUTPUT_DIR, --output-dir OUTPUT_DIR
                        Output directory to write results to.
  --save-plots SAVE_PLOTS
                        Whether to save plots.
  -m MASK, --mask MASK  Mask the image.
  -w WARNINGS, --warnings WARNINGS
                        Whether to ignore warnings.

Reducing Output#

If you find the output too verbose or of no use you can reduce it by setting the log_level to either error or warning. This can be done either in the configuration file (see Configuration below) or using the -l/--log-level flag for example topostats process --log_level warning.

Configuring TopoStats#

Configuration of TopoStats is done through a YAML file and a full description of the fields used can be found under the configuration section.

Here we will go through generating a configuration file to edit and some of the common changes that you are likely to want to make to the default configuration and how to make them.

Generating Configuration File#

TopoStats will use some reasonable default parameters by default, but typically you will want to customise the parameters that are used. This is achieved using a configuration file. This is a YAML file that contains parameters for different settings. For convenience you can generate a sample configuration file in your current working directory using the topostats create-config sub-command. It takes a single argument, the name of the file to save the configuration to (e.g. config.yaml or settings.yaml), and it will write the current default configuration to that file.

topostats create-config --filename my_config.yaml
ls -l

You can now edit and/or rename the my_config.yaml. It can be called anything you want, e.g. todays_first_run_configuration.yaml is a valid name.

Editing config.yaml#

IMPORTANT This file is an ASCII text file and you should use NotePad (Windows), TextEdit (OSX) or Nano/Emacs/Vim (GNU/Linux) or any other text editor. Do not use Microsoft Word or any other Word Processor to edit this file.

You can now start customising the configuration you are going to run TopoStats with. All fields have defaults but the ones you may want to change are….

  • base_dir (default: ./) the directory in which to search for scans. By default this is ./ which represents the directory from which topostats process is called and it is good practice to have one configuration file per batch of scans that are being processed.

  • output_dir (default: output) the location where the output is saved, by default this is the directory output which will be created if it doesn’t exist. If you wish for the output to be somewhere else specify it here. If you want Processed directories to sit within the directories that images are found then simply set the output_dir to the same value as base_dir.

  • log_level (default: info) the verbosity of output to the console and log file, the options in order of verbosity are debug > info > warning > error. If you want less output set to warning or error. If you encounter errors please set to debug and run again and include the log in your bug report.

  • cores (default: 2) the number of parallel processes to run processing of all found images. Set this to a maximum of one less than the number of cores on your computers CPU. If unsure leave as is, but chances are you can increase this to at least 4 quite safely.

  • file_ext (default: .spm) the file extension of scans to search for within the current directory. The default is .spm but other file format support is in the pipeline.

  • plotting : image_set (default core) specifies which steps of the processing to plot images of. The value all gets images for all stages, core saves only a subset of images.

Most of the other configuration options can be left on their default values for now. Once you have made any changes save the file and return to your terminal.

Running TopoStats with my_config.yaml#

To use your new configuration file you need to inform topostats process to use that file rather than the defaults, this is done using the --config config.yaml file.

NB this assumes that you are in the same directory as your scans where you have saved the my_config.yaml file that you edited. That doesn’t have to be the case but it makes life easier for if you are not familiar with absolute and relative paths.

topostats process --config my_config.yaml
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Configuration is valid.
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Plotting configuration is valid.
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Configuration file loaded from      : None
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Scanning for images in              : /home/neil/work/projects/topostats/TopoStats
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Output directory                    : output
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Looking for images with extension   : .spm
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Images with extension .spm in /home/neil/work/projects/topostats/TopoStats : 1
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Thresholding method (Filtering)     : std_dev
[Tue, 15 Nov 2022 12:39:48] [INFO    ] [topostats] Thresholding method (Grains)        : std_dev

On successful completion you should see the same message noted above.


The output from running TopoStats is saved in the location defined in the configuration file by output_dir. The default is the directory output within the directory from which topostats process. This may differ if you have used your own customised configuration file (specifically if you have modified the output_dir: option).

At the top level of the output directory are two files config.yaml and all_statistics.csv

  • config.yaml : a copy of the configuration used to process the images.

  • all_statistics.csv : a Comma Separated Variable ASCII plain-text file of the grain and DNA tracing statistics.

The remaining directories of results is contingent on the structure of files within the base_dir that is specified in the configuration. If all files are in the top-level directory (i.e. no nesting) then you will have just a Processed directory. If there is a nested structure then there will be a Processed directory in each folder that an image with the specified file_ext has been found. This is perhaps best illustrated by way of example.

If you have the following three .spm files within your current directory, one at the top level, one under level1 and one under level1/a

[4.0K Nov 15 13:55]  .
|-- [4.0K Nov 15 13:54]  ./level1
|   |-- [4.0K Nov 15 13:54]  ./level1/a
|   |-- [ 32M Nov 15 13:54]  ./level1/a/minicircle.spm
|   |-- [ 32M Nov 15 13:54]  ./level1/minicircle.spm
|-- [ 32M Nov 15 13:54]  ./minicircle.spm

…then under output (the default foroutput_dir) you will see the following directory structure.

[4.0K Nov 15 14:06]  output
|-- [ 381 Nov 15 14:06]  output/all_statistics.csv
|-- [7.4K Nov 15 14:06]  output/config.yaml
|-- [4.0K Nov 15 14:06]  output/level1
|   |-- [4.0K Nov 15 14:06]  output/level1/a
|   |   |-- [4.0K Nov 15 14:06]  output/level1/a/Processed
|   |-- [4.0K Nov 15 14:06]  output/level1/Processed
|-- [4.0K Nov 15 14:06]  output/Processed

…where there is one Processed directory at the sub-directory level that each image was found.

NB If you want Processed directories to sit within the directories that images are found then simply set the output_dir to the same value as base_dir.

Within each Processed directory is a directory for each file found with the specified file_ext and within these are the resulting images from processing scans. If the plotting : image_set is core then there is a single image for each. If this option is all then there is also a sub-directory for each image found within which there are the directories filters, grains/below and grains/above which contain additional images from the processing stages and an accompanying histogram for each image showing the distribution of pixel heights for that image.

Summary Plots#

By default TopoStats will take the data that has been summarised across all files and generate a series of plots, histograms with Kernel Density Estimates (KDE) overlaid and Violin plots. The default location of these if no custom configuration file is used is output/summary_distributions. If you have used a custom configuration file it will be the sub-directory summary_distributions nested under the directory specified for the output, e.g. if you used the current directory as output you will have a summary_distributions directory present.

Sometimes you may have a all_statistics.csv from a run and wish to plot distributions of additional statistics that were not already plotted. This can be achieved using the command line programme toposum which is included.

NB Because of the inherent complexity of plots this script is, by design, limited in the scope to which plots can be configured. It uses the plotting library Seaborn (which is built on top of Matplotlib) to produce basic plots, which are not intended for publication. If you want to tweak or customise plots it is recommended to load all_statistics.csv into a Jupyter Notebook and generate the plots you want there. A sample notebook is included to show how to do this.

Configuring Summary Plots#

Configuration of summary plots is also via a YAML configuration file a description of the fields can be found under configuration page. You can generate a sample configuration by invoking the --create-config-file option to toposum

toposum --create-config-file custom_summary_config.yaml

The file custom_summary_config.yaml can then be edited to change what plots are generated, where they are saved to and so forth. Typically you will only want to adjust a few settings such as toggling the types of plots (hist, kde and violin), the number of bins in a histogram or the statistic to plot in histograms (count, frequency etc.). You can change the palette that is used by Seaborn and crucially toggle which statistics are summarised by commenting and uncommenting the statistic names under stats_to_sum.


Labels for the plots are generated from the file topostats/var_to_label.yaml which provides a dictionary that maps the variable name as the dictionary key to its description stored in the dictionary value. If you wish to customise these you can do so and pass it to toposum using the --plotting_dictionary which takes as an argument the path to the file you have created.


The option pickle_plots: True will save to the specified output_dir the file distribution_plots.pkl which is a binary format that saves the plots that have been generated and saved in nested dictionaries so that they can be loaded again. The Notebook notebooks/02-Summary-statistics-and-plots.ipynb shows how to load these and make simple modifications to the the plots.