Introduction to image analysis using Galaxy

Authors: orcid logoAvatarThomas Wollmann orcid logoAvatarSaskia Hiltemann
Overview
Questions:

  • How do I use Galaxy with imaging data?

  • How do I convert images using Galaxy?

  • How do I display images in Galaxy?

  • How do I filter images in Galaxy?

  • How do I segment simple images in Galaxy?

Objectives:

  • How to handle images in Galaxy.

  • How to perform basic image processing in Galaxy.

Requirements:
Time estimation: 1 hour
Level: Introductory
Supporting Materials:
Last modification: Oct 18, 2022
License: Tutorial Content is licensed under Creative Commons Attribution 4.0 International License The GTN Framework is licensed under MIT

Introduction

Image analysis is the extraction of meaningful information from images by means of digital image processing techniques. Imaging is an important component in a wide range of scientific fields of study, such as astronomy, medicine, physics, biology, geography, chemistry, robotics, and industrial manufacturing.

This tutorial shows how to use Galaxy to perform basic image analysis tasks such as format conversion, image enhancement, segmentation, and feature extraction.

Agenda

In this tutorial, we will deal with:

  1. Introduction
  2. Getting data
  3. Image Metadata Extraction
  4. Image Conversion
  5. Image Filtering
  6. Segmentation
  7. Conclusion

Getting data

The dataset required for this tutorial is available from Zenodo and contains a screen of DAPI stained HeLa nuclei (more information). We will use a sample image from this dataset for training basic image processing skills in Galaxy.

Our objective is to automatically count the number of cells contained in this image. In order to achieve this, we will enhance the quality of the image, automatically detect the nuclei and segment the nuclei and count them.

Hands-on: Data upload

  1. If you are logged in, create a new history for this tutorial

    Click the icon at the top of the history panel.

    If the is missing:

    1. Click on the icon (History options) on the top of the history panel
    2. Select the option Create New from the menu
  2. Import the following dataset from Zenodo or from the data library (ask your instructor).
    • Important: Choose the type of data as zip.
    https://zenodo.org/record/3362976/files/B2.zip
    • Copy the link location

    • Open the Galaxy Upload Manager ( on the top-right of the tool panel)

    • Select Paste/Fetch Data

    • Paste the link into the text field

    • Press Start

    • Close the window

    As an alternative to uploading the data from a URL or your computer, the files may also have been made available from a shared data library:

    • Go into Shared data (top panel) then Data libraries
    • Navigate to the correct folder as indicated by your instructor
    • Select the desired files
    • Click on the To History button near the top and select as Datasets from the dropdown menu
    • In the pop-up window, select the history you want to import the files to (or create a new one)
    • Click on Import
  3. Unzip file with the following parameters:
    • “input_file”: Zipped input file
    • “Extract single file”: Single file
    • “Filepath”: B2--W00026--P00001--Z00000--T00000--dapi.tif

  4. Rename the dataset to input.tif

    • Click on the pencil icon for the dataset to edit its attributes
    • In the central panel, change the Name field
    • Click the Save button

Image Metadata Extraction

Now, we can extract metadata from an image.

Hands-on: Extract Image Metadata
  1. Image Info with the following parameters to extract metadata from the image:
    • “Input Image”: input.tif file (output of the previous step)

  2. Click on the (eye) icon next to the file name, to look at the file content and search for image acquisition information

    Question
    1. What is the datatype?
    2. What are the pixel dimentions?
    3. How many bits per pixel are used?
    1. TIFF
    2. 1344x1024
    3. 16

Image Conversion

Not all tools can handle all image formats. Especially proprietary microscope image formats should be converted to TIFF (supported formats). However, TIFF can not be displayed in the browser. Therefore, we convert input.tif to a PNG for visualization.

Hands-on: Convert Image
  1. Convert image with the following parameters to convert the image to PNG:
    • “Input Image”: input.tif file
    • “Output data type”: PNG
  2. Rename the generated file to viz_input
  3. Click on the (eye) icon next to the file name to look at the file content

Your image should look something like this:

raw input image.

Question

You can observe that the image content is barely visible. Why?

The original image is 16-bit and the intensity values are spread over a larger range than the display can render. Therefore, for improved visibility the intensity histogram of the image can be normalized first.

Next we will normalize the histogram to improve the contrast. We do this using a Contrast Limited Adaptive Histogram Equalization (CLAHE) approach.

Hands-on: Normalize Histogram and Convert Image
  1. Histogram equalization with the following parameters to normalize the histogram of the image:
    • “Source file”: input.tif file
    • “Histogram Equalization Algorithm”: CLAHE
  2. Rename the generated file to input_normalized
  3. Convert image with the following parameters to convert the image to PNG:
    • “Input Image”: input_normalized file (output of Histogram equalization )
    • “Output data type”: PNG
  4. Rename the generated file to viz_normalized
  5. Click on the (eye) icon next to the file name, to look at the file content

Your image should now look something like this:

viz_normalized output image.

We can now clearly make out the presence of the stained nuclei. Next we will automatically detect these features and segment the image.

Image Filtering

Specific features of interest (e.g., edges, noise) can be enhanced or suppressed by using an image filter.

Hands-on: Filter image
  1. Filter Image with the following parameters to smooth the image:
    • “Image type”: Gaussian Blur
    • “Radius/Sigma”: 3
    • “Source file”: input.tif file
  2. Rename the generated file to input_smoothed
  3. Histogram equalization with the following parameters to normalize the histogram of the image:
    • “Source file”: input_smoothed file (output of Filter image )
    • “Histogram Equalization Algorithm”: CLAHE
  4. Rename the generated file to input_smoothed_normalized
  5. Convert image with the following parameters to convert the image to PNG:
    • “Input Image”: input_smoothed_normalized file (output of Histogram equalization )
    • “Output data type”: PNG
  6. Rename the generated file to viz_smoothed_normalized
  7. Click on the (eye) icon next to the file name, to look at the file content and compare the result with viz_normalized. You can observe that viz_smoothed_normalized has significant reduced noise.

Your image should now look something like this:

viz_smoothed_normalized output image.

Segmentation

Objects of interest like nuclei can be segmented by using a smoothed image and thresholding. Moreover, the results can be overlayed with the original image.

Hands-on: Segment image
  1. Auto Threshold with the following parameters to segment the image:
    • “Source file”: input_smoothed file (output of Filter image )
    • “Threshold Algorithm”: Otsu
    • “Dark Background”: Yes
  2. Rename the generated file to input_segmented
  3. Binary 2 Label with the following parameters to segment the image:
    • “Binary Image File”: input_segmented file (output of Auto Threshold )
  4. Rename the generated file to input_segmented_labeled
  5. Convert image with the following parameters to convert the image to PNG:
    • “Input Image”: input_segmented_labeled file (output of Binary 2 Label )
    • “Output data type”: PNG

  6. Rename the converted image to viz segmented

    Question
    1. What does Binary 2 Label do? (Hint: check the tool help section)
    2. View the viz_segmented image from the last step, what do you see?
      • Can you explain this result?
    3. Exercise: Try to make the information in this image better visible (Hint: Histogram Equalization )
    1. The tool assigns each connected component (e.g., segmented cell) in the image an object id and stores it as the intensity value.
    2. The image looks completely black. The object IDs generated by Binary 2 Label are relatively low. Since the IDs are stored as intensity values, these are too low to be visible in this case. Nevertheless, there is more information in this image than meets the eye.

    3. To make labeled objects visible, the values have to be stretched to a larger range of visible intensity values. We can do that by equalizing the histogram again:

      • Histogram equalization with the following parameters to normalize the intensity values:
        • “Source file”: input_segmented_labeled file (output of Binary 2 Label )
        • “Histogram Equalization Algorithm”: CLAHE
      • Convert image with the following parameters to convert the image to PNG:
        • “Input Image”: output of Histogram Equalization
        • “Output data type”: PNG

      The information contained in the original image has now become visible to the human eye: normalized viz_segmented file.

  7. Overlay Segmentation Mask with the following parameters to convert the image to PNG:
    • “Image Source File”: viz_normalized file
    • “Mask Source File”: viz_segmented file
    • “Image Is Greyscale”: Yes
    • “Thickness”: 3
    • “Stroke Color”: red
    • “Plot Labels”: yes
    • “Label Color”: yellow
  8. Click on the (eye) icon next to the file name, to look at the file content and assess the segmentation performance
  9. Count Objects with the following parameters to count the segmented objects in the image:
    • “Source file”: input_segmented_labeled file (output of Binary 2 Label )
    Question

    How many objects were segmented?

    The Count Objects tool counted 425 objects.

The resulting image should look something like this:

segmentation mask output image.

We see the segmentation mask overlayed; each detected object (nucleus) is labeled with its ID value.

We see that with the help of just a few simple steps, we were able to detect the locations of the stained nuclei, and count them.

Conclusion

In this exercise you imported images into Galaxy, extracted meta information from an image, converted between file formats, learned how to visualize microscopy images, filtered the image, and segmented cells using Galaxy.

Key points

  • The Image Info tool can provide valuable metadata information of an image.

  • TIFF files can not directly viewed in the browser, but have to be converted.

  • For visualization, images with a bit-depth more than 8-bit have to be histogram equalized.

Frequently Asked Questions

Have questions about this tutorial? Check out the tutorial FAQ page or the FAQ page for the Imaging topic to see if your question is listed there. If not, please ask your question on the GTN Gitter Channel or the Galaxy Help Forum

Useful literature

Further information, including links to documentation and original publications, regarding the tools, analysis techniques and the interpretation of results described in this tutorial can be found here.

Feedback

Did you use this material as an instructor? Feel free to give us feedback on how it went.
Did you use this material as a learner or student? Click the form below to leave feedback.

Click here to load Google feedback frame

Citing this Tutorial

  1. Thomas Wollmann, Saskia Hiltemann, 2022 Introduction to image analysis using Galaxy (Galaxy Training Materials). https://training.galaxyproject.org/training-material/topics/imaging/tutorials/imaging-introduction/tutorial.html Online; accessed TODAY
  2. Batut et al., 2018 Community-Driven Data Analysis Training for Biology Cell Systems 10.1016/j.cels.2018.05.012 


@misc{imaging-imaging-introduction,
author = "Thomas Wollmann and Saskia Hiltemann",
title = "Introduction to image analysis using Galaxy (Galaxy Training Materials)",
year = "2022",
month = "10",
day = "18"
url = "\url{https://training.galaxyproject.org/training-material/topics/imaging/tutorials/imaging-introduction/tutorial.html}",
note = "[Online; accessed TODAY]"
}
@article{Batut_2018,
    doi = {10.1016/j.cels.2018.05.012},
    url = {https://doi.org/10.1016%2Fj.cels.2018.05.012},
    year = 2018,
    month = {jun},
    publisher = {Elsevier {BV}},
    volume = {6},
    number = {6},
    pages = {752--758.e1},
    author = {B{\'{e}}r{\'{e}}nice Batut and Saskia Hiltemann and Andrea Bagnacani and Dannon Baker and Vivek Bhardwaj and Clemens Blank and Anthony Bretaudeau and Loraine Brillet-Gu{\'{e}}guen and Martin {\v{C}}ech and John Chilton and Dave Clements and Olivia Doppelt-Azeroual and Anika Erxleben and Mallory Ann Freeberg and Simon Gladman and Youri Hoogstrate and Hans-Rudolf Hotz and Torsten Houwaart and Pratik Jagtap and Delphine Larivi{\`{e}}re and Gildas Le Corguill{\'{e}} and Thomas Manke and Fabien Mareuil and Fidel Ram{\'{\i}}rez and Devon Ryan and Florian Christoph Sigloch and Nicola Soranzo and Joachim Wolff and Pavankumar Videm and Markus Wolfien and Aisanjiang Wubuli and Dilmurat Yusuf and James Taylor and Rolf Backofen and Anton Nekrutenko and Björn Grüning},
    title = {Community-Driven Data Analysis Training for Biology},
    journal = {Cell Systems}
}

Congratulations on successfully completing this tutorial!