“The virtual microscope helps to analyse micro-organisms faster and better”
In this series of interviews we show what contribution projects can make to FAIR research IT. The research teams of the projects have received a grant from the FAIR Research IT Innovation Fund.
Studying micro-organisms such as pollen grains under a virtual microscope, can it be done? Yes, soon it will be. In the MicroMap project a large pollen collection is digitised and made public via an online and interactive environment.
This gives researchers and students around the world opportunities for a faster and better (re)-analysis, training and documentation. What is more, the user-friendly software also allows you to study other kinds of micro-organisms. Initiator Timme Donders explains how the virtual microscope works and how the project contributes to working FAIR and open science.
As a “pollen researcher” Timme Donders has already spent many hours behind a microscope. He is a biologist by background and within the department of Physical Geography researches vegetation change. “How has the landscape evolved? How does this affect ecosystems? I try to answer such questions by analysing fossil pollen grains under the microscope. What pollen types does it contain? In what proportions? In this way I make reconstructions of the climate and the landscape, going back millions of years. This information is used, for instance, to improve climate models.”
Time-consuming analyses
So important research, but the analysis process is not ideal. ‘ Researchers like me spend 90 % of the analysis time on identifying common issues. For instance, a specimen is often full of birch pollen. Not that exciting, but it has to be counted nevertheless.” In addition, it takes a lot of effort to learn someone to analyse the micro-organisms properly “ After a microscopic analysis I do not exactly know what a student has done. That makes it tricky to assess someone’s knowledge and skills.” Add to this the fact that the large pollen collection at Utrecht ľϸӰ is accessible to only a few people. “That’s a shame, because students and researchers all over the world could use it to do interesting analyses.”
Collaboration with UMCU
To tackle these problems, Donders toyed with the idea of a virtual microscope making scanned specimens visible and analysable in an online environment. “But I had no expertise on image-analysis. Fortunately, Utrecht ľϸӰ organized a pioneer project on the theme of imaging. All kinds of research groups came together. I signed up after which the project coordinator referred me to a group in the UMCU specialising in microscopy and automatic object identification. This group works with large quantities of image data, from X-rays to images from pathology. So no pollen, but their techniques I could use.” To combine their expertise, Donders started a collaboration with the group. Meanwhile, he had the FAIR Research IT Innovation Fund in mind. With input from the UMCU-programmers (Edwin Bennink and Christian Mol) he wrote a proposal for the MicroMap project. The application was granted, after which Donders as project coordinator put together a permanent team with the UMCU group.
Digital pollen database
First of all, the physical pollen collection is digitised. “ We scan in all the specimens. This provides an online database containing pollen of identified plant species. Now more people around the world can have a look at it. This makes training and looking up information much easier. In addition, we record the data sustainably, because such a pollen collection decays slowly.” The digital reference collection is especially interesting for people working with pollen grains. “Researchers like me, but also analysts from Pollenradar, for example, benefit from it. It is specific knowledge that requires a lot of training.” You can also work with specimens that have not yet been scanned. “After some explaining it takes about thirty minutes to digitise a specimen with our microscope. That is peanuts compared to the time taken up by the old analysis.”
Virtual microscope
Parallel to the digital collection, Donders’ team is building a virtual microscope. This consists partly of software to smartly visualise microscopic image data (“Pollenlabeler”) and partly of a virtual environment where people can analyse those images (MicroMap). “ We fill our database with the images of the scanned pollen. We are also building a tool to display all the image data properly. Challenging, because detailed 3D scans quickly become huge. The tool cuts out only the interesting objects. Someone can look at these images in detail and zoom through them in 3D, as if you were behind a microscope. As a result, anybody can analyse the existing collection via a freely accessible webpage.” This is especially useful to train people, the researcher says. Both for self-study – students learn to correctly classify the elements – and supervision. “It improves educational quality and offers options for online teaching. When students do their analyses with the Pollenlabeler you can see exactly what they have identified. And if they don’t know a category? Then you can easily check it out together.They can share their final results via MicroMap.”
Broadly applicable tool for analysis and training
Do researchers working with micro-organisms other than fossils also benefit from the project? Yes, they do, because the team shares the software they use to fill the database and create the online environment. “ Although we are building a virtual environment around the scanned pollen collection, the underlying code can also be used to study and digitise other organisms. The code is used to read in images from the microscope and put into the right format for a virtual environment.” This tool is interesting for anyone who analyses micro-organisms or wants to train people on this topic. “In science and practice, complex, 3D-ish images of micro-organisms are widely used. Think for instance of algae-analyses as part of monitoring water quality. Or blood cell analyses in the medical world. MicroMap is an excellent solution for such applications, making the analyses faster, more effective and better reproducible. It is also of interest to fields such as forensic research, archeology and industry. I am curious to know who will be showing their interest.” In the near future anyone will be able to download the Pollenlabeler code, but for now it can only be done via the paid programming environment Matlab (free for UU-users). In the future, the team wants to convert the code to a fully open format, such as Python. Furthermore, the tool has been made as user-friendly as possible, so without complicated codes. You can also use different image formats.
From monitoring water quality to blood cell analysis: for all kinds of applications, MicroMap is a great solution that makes microscopic analyses faster, more effective and better reproducible.
Better reproducible
In short, in many ways MicroMap contributes to FAIR working and the open science ambitions of Utrecht ľϸӰ. Donders: “It also makes the analysis results of micro-organisms better reproducible. Now data is shared based on derived data, such as identifications. But with MicroMap you share the original images. As a result, others can verify your results.” It also improves the comparability of the results between researchers. “By digitising data and sharing it centrally, you can better agree on identifications. It makes (international) collaboration easier.”
The fund as an accelerator
How does Donders look at the “journey” so far? “To realise an idea like this, you have to enjoy looking beyond what you are used to. Don’t hold yourself back, just because you do not understand something. By asking lots of questions and seeking out the right people, I learn more and more.” In his discipline there has long been a desire to automate the analysis process and to share data, he says. “But you need the right resources and people. Now the budget allowed me to hire the UMCU programmers, which would not have been possible otherwise. I was also able to hire a student, Romee van der Kuil, who scanned pollen and helped develop the tool with the help of her programming skills.” Initiatives such as the pioneer project around imaging and the FAIR fund make it possible to realise these kinds of ideas, Donders says. “The project plan and the deadlines also provide structure. Without this fund, the idea would probably not have got off the ground. I am grateful for this opportunity.”
By asking a lot of questions and seeking out the right people, I am slowly learning more and more.
Future
Finally: how far along are Donders and his team? “Within the project we can digitise approximately half of our pollen. The rest will follow later. Hopefully, the tool and the online environment will go live before 2025.” He is already noticing interest from various quarters, from geological services and forensic institutes to archaeological research agencies. “Originally my field of research is not strongly associated with this kind of digital research techniques. So when you develop something like this, it attracts attention. Especially because the world of microscopy is interested in automatic detection of objects. This automation we are also trying to pursue with a number of partners. That is a larger process than this project, but the platform we are setting up now provides a steady basis. By smart collaboration, we are taking science forward step by step.”
Probably the virtual environment with the pollen collection can soon be viewed via:
About the FAIR Research IT Innovation Fund
Utrecht ľϸӰ wants every research team to be well supported in the field of research IT. That is why there is the FAIR Research IT Innovation Fund. Think of advice and support in ensuring that there is enough storage capacity for data. Or the development of tools and services that help researchers in their work. When selecting projects the FAIR and open science principles form the guideline. For instance, other researchers must be able to easily reuse the knowledge and solutions. Do you have questions? Please contact us.