Skip to main content
Publications of the Week

Label-Free Prediction of Three-Dimensional Fluorescence Images from Transmitted-Light Microscopy

By October 1, 2018No Comments

Read the Publication

 This week we profile a recent publication in Nature Methods from
Dr. Gregory Johnson at the Allen Institute for Cell Science (pictured).

Can you provide a brief overview of your lab’s current research focus?

The mission of the Allen Institute for Cell Science is to create dynamic and multi-scale visual models of cell organization, dynamics and activities that capture experimental observation, theory and prediction to understand and predict cellular behavior in its normal, regenerative, and pathological contexts. We have started with imaging the major components of the cell. To do this we primarily use two technologies, CRISPR, a type of “molecular scissors” to create different types of cells (we call these cell lines), each with a code for a fluorescent molecule attached to a distinct cell component (an organelle), and with fluorescence microscopy, which uses special lasers that causes the fluorescence molecules to light up and to allow us to take images of these specific components. These images are extraordinarily complex, and tell us a lot about the natural variation in cell biology and how cells reorganize as they turn from stem cells to other cell types such as muscle cells, kidney cells, etc. One of the major difficulties of building these models of cell organization is figuring out how to integrate these images of different components into a coherent model of how these parts fit together.

What is the significance of the findings in this publication?

Building these cell lines is a laborious task – we have multiple upstream teams dedicated to their creation. Because stem cells are so fragile, taking fluorescence images – which requires shining laser light on the cells – may disturb or damage them. This prevents us from imaging many cellular components at the same time and from making longer observations about cell behaviors There are microscopy imaging methods that are easier, less expensive, and gentler for cells than fluorescence microscopy, such as shining a normal white-light through the cells and taking a picture, called transmitted light images. Unfortunately, these images can be very difficult to interpret, and show everything in the cell with no way to distinguish one cellular structure from another, preventing us from understanding how different cell components interact and organize. We found that if we take transmitted light images and fluorescence images of the same cells at the same time, we can build a model to predict, with high accuracy, the fluorescence image from the transmitted light image. If we build a different model for each cell component we’re interested in, we can take cells without the fluorescent label and use transmitted light microscopy to determine where many of the major components of the cells are, how they are organized with respect to each other, and how that organization changes over time.  This saves us an incredible amount of time and effort, and it allows us to see things that we never would have been able to using traditional cell-imaging methods. Furthermore, the model building is straightforward enough that other scientists can use our tools and build models for their research as well.

What are the next steps for this research?

We want to build bigger, more accurate models that allow us to see at higher resolution. We want to identify and explore how important components reorganize as cells mature and grow. We want to see where the model works and where it doesn’t. We are giving our images and models away on our website to allow other scientists to try their ideas and pursue new discoveries. This research opens up tons of possibilities that we are just beginning to explore. The hardest part is figuring out where to spend our time!

This research was funded by:

This work was supported by grants from NIH/NINDS (R01NS092474) and NIH/NIMH (R01MH104227).
We also thank Paul G. Allen, founder of the Allen Institute for Cell Science, for his vision, encouragement and support.

Read the Publication