Running PIE single-image analysis

To analyze a single image using PIE, you can use the analyze_single_image function.

Inputs

• input_im_path: {path}

  the path to the image to be analyzed.

• output_path: {path}

  the directory in which image analysis folders should be created

• image_type: {‘bright’ or ‘dark’}

  The brightness of the cell bodies relative to the image background.

• hole_fill_area: {integer or ‘inf’}, optional, default ‘inf’

  the area (in pixels) of the largest size hole to fill in colony masks after image analysis. For low-resolution yeast imaging, we recommend setting this value to ‘inf’ (i.e. all the holes in the colonies get filled)

• cleanup: {True or False} optional, default False

  whether or not to perform ‘cleanup’ of spurious pieces of background attached to colonies. (We recommend trying PIE both with and without cleanup on a set of sample images; you can see the Li, Plavskin et al. paper for details)

• max_proportion_exposed_edge: {0-1} optional, default 1

  maximum proportion of the perimeter of a PIE-detected gradient object (‘PIE piece’) that may be non-adjacent to another PIE piece to avoid being removed during ‘cleanup’ steps; only used if cleanup is True. 0.75 works well for 10x yeast images, and many other images tried

• cell_intensity_num: {1 or 2} optional, default 1

  number of distinct intensity categories belonging to cells in the image; determines whether 2 or 3 Gaussians are used to fit intensity log histogram for threshold determination. Can be set to 1 (default) or (in rare cases) 2.

• save_extra_info: {True or False} optional, default True

  whether to write additional files described above

command-linepython

pie analyze_single_image INPUT_IM_PATH OUTPUT_PATH IMAGE_TYPE

or, with options

pie analyze_single_image INPUT_IM_PATH OUTPUT_PATH IMAGE_TYPE -h HOLE_FILL_AREA -c CLEANUP -m MAX_PROPORTION_EXPOSED_EDGE -s SAVE_EXTRA_INFO

Note

Inputs besides INPUT_IM_PATH, OUTPUT_PATH, and IMAGE_TYPE are optional.

Windows cmd Examples

Analyze an image with bright cells on a dark background, E:\trial\my_image.tif, and save outputs in a new folder called E:\trial_output_images

pie analyze_single_image E:\\trial\\t01xy0001.tif E:\\trial_output_images bright

or, with cleanup set to True and hole_fill_area set to 40:

pie analyze_single_image E:\\trial\\t01xy0001.tif E:\\trial_output_images bright -h 40 -c True

Caution

You will need to use a double backslash (’\\’) instead of any backslash (’\’) characters in any file paths.

If your filepath has a space in it, you will need to surround the path name with quotation marks, e.g.:

pie analyze_single_image E:\\trial\\t01xy0001.tif "E:\\my output folder" bright

MacOS/Unix Terminal Examples

Analyze an image with bright cells on a dark background, ~/trial/my_image.tif, and save outputs in a new folder called ~/trial_output_images

pie analyze_single_image ~/trial/t01xy0001.tif ~/trial_output_images bright

or, with cleanup set to True and hole_fill_area set to 40:

pie analyze_single_image ~/trial/t01xy0001.tif ~/trial_output_images bright -h 40 -c True

Caution

If your filepath has a space in it, you will need to prepend the space with ‘\’, e.g.:

pie analyze_single_image ~/trial/t01xy0001.tif ~/my\ output\ folder bright

Outputs

This function runs PIE, creates output folders within output_path, and writes files to:

• colony_masks:

  the colony mask, with each colony labeled in a different number, as a tif file

• jpegGRimages:

  a jpeg of the original image

• single_image_colony_centers:

  a csv file containing the properties (e.g. area) of all the colonies in the image.

If save_extra_info is True (default), then additional files are saved in the following folders:

• boundary_ims:

  a jpeg of the original image, overlaid with the contours of the colony mask

• threshold_plots:

  plots demonstrating the detection of the threshold based on the log histogram of a background-corrected image, and files with information on curve fits and threshold values for thresholding

• colony_center_overlays:

  a jpeg of the original image, overlaid with the contours of the colony mask and a transparent mask of the cell centers detected after thresholding