Automated Zebrafish screening for Leukemia drug discovery

In this interview,  Elspeth Payne and Alexandra Lubin, cancer researchers at the UCL Cancer Institute, talk about how they use zebrafish modeling to study myeloid leukemias in a high-throughput, automated fashion using IDEA Bio-Medical’s AI-powered analysis software, Athena. This affordable software solution is available for download in a novel pay-per-use basis.
Could you please introduce yourself, what it is you do and briefly describe your research?

Elspeth Payne: My name is Beth Payne and I am a clinician scientist. I work on blood cancers in the laboratory and look after patients in the hospital. My laboratory works on, particularly, myeloid malignancies. We use a combination of zebrafish modeling and drug screens alongside primary patient material to try and study better ways to understand the development of myeloid leukemias and different ways that we could treat them in the future.

Alexandra Lubin: My name is Alex Lubin. I am a post-doctoral associate in Beth Payne’s lab. I work on zebrafish models of myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). Drug screening is my main project and I am looking at doing some drug screening in hematopoietic stem cells in the tails of zebrafish to look for synthetic lethality.

How did you perform analysis of zebrafish before using the Athena software?

Alexandra Lubin: When I am looking at stem cells in my drug screen, I try to find a change in number of cells. To do this manually, all the individual fluorescent cells must be counted in the tail. That means you have to take each fish under a microscope, go through and count them, which is very time-consuming. For this reason, we started looking for an automated approach.

What are some of the challenges you face in getting the data you need?

Beth Payne: We have been developing some drug screening methodologies, and while the zebrafish is a great model for this, getting the right throughput has been challenging. Although there are many different platforms that one can use, they tend to have kind of bespoke solutions for how you analyze the data. That means it is costly and takes a long time to develop a process that works for your purpose.

At what point then did you move away from manual imaging and image analysis to search for automated solutions?

Beth Payne: In terms of automated screening, we have been looking for a solution for probably the last five to ten years. Every few years we revisited this as a key goal, asking “how can we make the process more time efficient.”  Manually imaging and counting the cells was the only way to quantify your data.  But we knew if it could be automated, firstly you take away all the time dedicated to doing those activities, but it also opens whole host of other opportunities to use the platform for other types of experiments.

What are the main challenges when sampling with a manual workflow?

Alexandra Lubin: The main challenge with the manual method is that it is time-consuming. It is not particularly difficult to do, but it involves spending hours and hours at microscopes. The other main problem with that is you cannot achieve large sample numbers. When you want to run something like a drug screen, the power comes from being able to analyze many samples.

What has this technology brought to your lab?

Beth Payne: The thing that has been helpful for this system is we use it for all indications rather than just drug screening. For example, if we have a transgenic or an in-situ that we want to image and look at in detail with high content, we will use the Athena rather than doing it manually.

The other great thing about the system is that it is flexible and you can optimize it for a different assay relatively quickly as the system is user-friendly. The company offers a lot of support to try and help us when we run into some difficulties.

I would say that overall, it allows us to power our experiments much better because instead of using just 20 embryos, we can use 200 for the same processing time. Overall, it offers more flexibility.

High-Content Analysis vs. Standard Analysis. A) Image showing fluorescence channels of a 4 dpf zebrafish embryo labeled with CD41:GFP [26] (orange) and lysC:mCherry [27] (cyan) to visualize hematopoietic stem cells and thrombocytes, and myeloid cells, respectively. Zoom-in highlights the cell types visible in the tail region. B) Brightfield and fluorescence channel overlay of the image in (A). C) Image from (A) with CD41:GFP and lysC:mCherry expressing cells identified by fluorescence image analysis to permit counting of cells or cell clusters. D) Image from (B) with zebrafish anatomy identified through automatic AI based image analysis [28] to group the identified cells/cell clusters found through fluorescence image analysis. E) Quantification of cell/cluster counting in the two fluorescence channels from panel (C). F) Quantification as in (E), but with each cell/cluster associated with the relevant zebrafish anatomy identified in (D), which also quantifies area of anatomical structures. Outline colors: Fish contour (yellow), head (blue), trunk (dark purple), tail (dark red), eye (orange), heart (dark pink), otic vesicle (golden), yolk sac (white), swim bladder (light pink), notochord (light green), tail fin (light blue), CD41:GFP (red spot outlines), lysC:mCherry (dark blue spot outlines). Image source: Technologies bringing young zebrafish from a niche field to the limelight Jason John Otterstrom, Alexandra Lubin, Elspeth Payne, Yael Paran, SLAS Technology, 2022, ISSN 2472-6303,
When you started working on the Athena software, how long did it take you to be independent and productive with it?

Alexandra Lubin: When I first encountered the Athena software, it was very much in development. That gave room for many discussions with IDEA Bio-Medical about what we wanted the software to do, what would be useful, and highlighting areas for improvement.

Once it was up and running, the transition time from first using the software to becoming fully independent with it was relatively short.

How does that compare with other instrumentation you use, for example compared with manual microscopes?

Alexandra Lubin: This has been a unique experience in the sense that we have had more contact with IDEA Bio-Medical than we would usually have. I have never used another system like this, so I do not really have a direct comparison, but it is an excellent and extremely practical piece of software.

How long does it take you to analyze all your fish samples, and how does that compare previously?

Alexandra Lubin: When conducting these experiments manually, we can only screen about a hundred fish at a time. However, that takes such a significant amount of time, but once we began using the Athena, we could increase that number.

How many fish I screen depends on how well-behaved the fish are, as anybody who works with zebrafish will know. (In our lab) it is possible to screen hundreds (of fish) in a short space of time as it only takes around 30 minutes between loading the plate and getting the results for about a hundred fish, whereas previously, that would have taken a whole day.

What do you do when you need some support using the software? Who supports you and how?

Alexandra Lubin: When I need some support with the software, I reach out to the team at IDEA Bio-Medical. They have been really helpful and offer a collaborative partnership. I provide them with images; they provide us with support – it works seamlessly.

What was the point when you thought the Athena software was useful to your research?

Alexandra Lubin: When I started this drug screen, we knew we needed to bring in automation as it was not going to be possible to screen thousands of compounds manually. When we started looking at different methods, it became clear how easy to use the system was and how flexible it was. Moreover, we were drawn to the fact that Athena enabled us to branch out and change it if we needed to.

Where do you see your work going next and the impact?

Beth Payne: I believe that it has certainly opened the door to performing more and more screens. Perhaps we may now be able to conduct more ambitious types of screens where we might use, for example, more than one fluorophore or more than one genotype at the same time because the time taken to screen is so much shorter and more efficient that you can introduce more variables.  

IDEA Bio-Medical currently focuses on empowering zebrafish researchers, specifically, to provide them with a reliable, robust solution for automated and unbiased Zebrafish image analysis by applying company’s knowledge and expertise.

To this end, IDEA developed a novel deep learning-based image analysis software for in vivo zebrafish experiments. The software automatically detects zebrafish contour and their internal organs in brightfield with no required user input. The anatomy identified is coupled to fluorescence channels to permit anatomy-specific study of fluorescence changes. It is an affordable, user-friendly system designed specifically for reliable, automated zebrafish image-based analysis.

The software is available as a stand-alone product and accepts microscopy images in multiple image formats, including proprietary ones. It is suited for researchers who only image and analyze a handful of fish per week, as well as researchers imaging hundreds and thousands of fish in multi-well plates for large scale screens. IDEA Bio-Medical is offering a novel pay-per-use model to access the software to enable flexible access.  So, all researchers using manual microscopes or automated systems from other vendors, can readily use IDEA’s Zfish software to extract quantitative, meaningful information from their Zebrafish images when they need it.

About The Speakers

About Dr Elspeth Payne

Dr Elspeth Payne is a Senior Clinical Researcher/Clinical Consultant UCL Cancer Institute. Her laboratory at UCL’s Cancer Institute is dedicated to the study of inherited bone marrow failure disorders and leukemias, and uses zebrafish to model these diseases. Dr Payne is also a clinical hematologist at UCL Hospital, where she treats people with blood disorders including leukemia and bone marrow failure.

Payne Lab UCL Projects (

About Dr Alexandra Lubin

Dr. Lubin is a post-doctoral at the UCL Cancer Institute where she uses zebrafish to study myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML), aiming to develop novel therapeutic treatments. Previously, she obtained her PhD in Chemical Biology from Imperial College London after studying Chemistry at the University of Cambridge.