Microbiology software in action: A best practice case study
Find out how automated microbiology image analysis software revolutionizes fungal genome research. Learn about the latest applications of our AI-powered biological image analysis software in the field of applied microbiology. Expert microbiologists share their experience with microscopy image analysis of fungal cultures using the IKOSA Platform. Meet Prof. Irina Druzhinina, a leading researcher in the field of microbiology and fungal genomics, Fungal Genomics Group, at the Nanjing Agricultural University. She and her team throw light on advanced analytical methods and microscopy-based techniques in fungal research using our cutting-edge microscopic image analysis software.
The start of a fruitful cooperation in the field of microbiology research
Polina: Dear Irina, our cooperation with your research group started in 2018. At that time, we were a young startup developing the idea of a platform for automated microscopy- or other biological image analysis software using applications based on artificial intelligence and machine learning.
You were the decision-maker to start the collaboration with us and also, one of our first clients who trusted in our solution. What potential benefits have you seen in our cooperation?
Irina: That time I still ran my group on Microbiology and Applied Genomics in ICEBE in TU Vienna. I was contacted by two young scientists or developers who proposed their unique method of analyzing biological images . And this idea truly attracted me because I didn't know any alternative or other company or service that could do it. And, of course, I was immediately fascinated by the opportunity to try an artificial intelligence platform for this purpose.
“I must say that it was particularly interesting for me to experience automated image analysis because I instantly saw the chance to develop something new for our field, which didn't exist before.”
Polina: And as I understand it’s related to your research work in fungal biology. Our collaboration did result in an article published in the top journal in the area of microbial ecology. Namely, Evolutionary compromises in fungal fitness. Our company actually developed a customized algorithm solution suited to the needs of your research group. Am I right?
Irina: Yes, absolutely correct.
Customizable microbiology software solutions for the purposes of fungal assay image analysis
Polina: As it’s stated in one of the articles, the main idea of it was to “automatically quantify the aerial mycelia and conidia abundance (% coverage) on each carbon source at the different time points measured. It was labelled as the REproduction Potential Artificial INTelligence assay (REPAINT) applicable for conidiation.” Could you tell a bit about the research and explain the benefits of using this algorithm? Please give our readers some further information, which product features of our software have proven the most useful when conducting your study. How did our individual solution help you improve the efficiency of your research?
Irina: We investigate the biology of a remarkable organism – the mold Trichoderma. It can make enzymes to degrade plant biomass which is crucially important for biofuel production, textile or paper industry. This fungus can also parasitize other fungi and therefore can be used for plant protection in agriculture. We need to understand the physiology of this fungus for the development of safe products for industry and biofungicides for agriculture.
Some Trichoderma species use air as the main medium for the dispersal of their spores and produce fluffy colonies (left), while others disperse by means of water or insects and therefore have almost no hyphae on the surface (right).
For the past 20 years, we have used a system called Phenotype MicroArrays. The custom-made microplates allow simultaneous monitoring of fungal growth on 96 carbon sources. The system was developed for bacteria by Biolog company in California, and our group in Vienna in collaboration with ECORC (Eastern Cereal and Oilseed Research Centre in Canada) was among the first users and developers of this system for industrially-relevant fungi such as the cellulase producing and plant-beneficial Trichoderma.
At a certain moment during this study, we noticed that fungi appeared different on the surface of these plates and started to take pictures along with estimating the growth by spectrophotometric measurements of turbidity. Even though we have used these images for several studies, their analysis remained qualitative and supplementary.
As data accumulated, we became more and more certain that the images tell us a lot about fungal development in addition to what we can measure. Some fungi rapidly formed spores on the surface of the plate while others developed slowly. And all these processes were influenced by carbon metabolism. But the evaluation was based on manual assessment of abundant images and therefore was very approximate and not quantifiable.
Trichoderma harzianum that is used for plant protection and plant growth promotion forms spores in response to nutritional limitation
When your company contacted me in 2018, that was the first thing I thought of: I was hoping that we could now analyze and quantify this data.
So, what we have done with the KML Vision team is the development of Phenotype MicroArrays to assess fungal reproduction potential, i.e. the sporulation.
“With the help of the algorithm, Fungal REPAINT Assay, we have a new level of understanding of biological information, which we can now easily collect from the same data.”
The results we are getting now with the REPAINT algorithm are incomparable to what we had before when we visually estimated the images. Your algorithm allows us to transform our images into excel tables of numerical data that can be statistically analyzed and tested. If previously we could just say if there was a difference or not, now we can really see the dynamics of fungal development, reproduction and death. This is a great advantage for our research on fungal fitness.
REPAINT microbiology software analysis output
REPAINT allows the statistical analysis of fungal conidiation and the colony forming using a great variety of conditions
“The results we are getting now with your algorithm are incomparable to what we had before when we visually estimated the images.”
Spores clamps of Trichoderma guizhouense ready for dispersal. Image credit: Feng Cai, www.FungiG.org
Polina: These words are the greatest motivation for us to continue the development of new algorithms dedicated to the field of microbiology! Thank you!
Irina: Actually, about the development. So far, we have published one study on REPAINT. But my group is actively involved in a large international project on Trichoderma genomics. This project is funded by the United States Department of Energy's Joint Genome Institute in California. And there we sequence a huge set of Trichoderma genomes. In this study, we of course also use the REPAINT Assay as one of the phenotypic characteristics.
So, this means that the study we have previously published was only the first introduction of this method. It now becomes one of the reliable tools and we want to expand its use and investigate more industrial fungi that play their roles for humankind. And we already have some ideas for the existing algorithm, but also its potential improvement and optimization.
“Since this method has performed well in previous research, I would like to add it in the tool box of standard assays for fungal phenotyping.”
Polina: We will be glad to provide the necessary algorithm optimization for your successful research. This is certainly an important study and we are grateful to be a part of it.
I have one question in regards to data transfer. Usually, during the research, scientists move around the world. As far as I know, in the course of this project, you also moved from Austria, the Technical University of Vienna to China, Nanjing Agricultural University.
How did data transfer management and coordinating the research process between the project participants work out for you?
Irina: At that moment, in 2018, I was still running my group in Vienna and my current assistant Feng Cai was already in China, where the majority of data was produced. Initially, by then there was no batch image analysis integrated in the IKOSA Platform, and we sent the data straight to your company. It was also convenient and very fast. However, now we manage it successfully within IKOSA by ourselves.
“And to the truth, no matter where we move to, we will continue using your IKOSA platform.”
Leading microbiologists on the ikosa software unique product features
Polina: That's great to hear! We can also assure you that, regardless of the physical location of your team members, our solution will always be accessible to everyone on the team to support research and access information in real time. And actually, one interesting fact in this regard.
As far as I know, when we first got in touch, you didn't know that AI-based image analysis software could somehow advance your microbiology research project. Now the market is getting saturated with other software providers specializing in this area and you may have already heard of some of them. What makes our solutions so distinct from those alternative products and why would you recommend working with us?
Irina: From my perspective, communication was an important point. I always got quick answers from your company, and your team is perfect at managing projects and work in general. It was a very pleasant, positive and fruitful experience. So... why should I change? I have not a single reason for that. We are totally satisfied.
And even more, if Michael remembers, at first, we had two ideas. First, I thought about analyzing electron microscopy images, which we also have in large numbers. However, then we decided to go for the analysis of Phenotype MicroArrays. As far as I know, today your company has successfully completed an AI algorithm for the analysis of electron microscopy images of mouse myocardium cells. It might be that this experience can also be applied for the analysis of fungal ultrastructure.
Transmission electron microscopy (TEM) image of a murine myocardial tissue section. The AI application automatically analyses the ultra-structure of the tissue. This enables morphometry of individual lipid droplets (blue), mitochondria (red), sarcomeres (green) and z-stripes (purple). Additionally, the distance between z-stripes are measured (green lines).
I remember we got a very reasonable and quick answer that you could try to run this analysis. And then again in a very short time, we received results. It was surprising! And for our project it offered an opportunity to move forward, so we immediately used it. In practical terms, I have no problems and, as a result, I am not interested in other solutions.
“In a very short time, we received results. It was surprising!”
Polina: Thank you, Irina, for your time and attention. We really appreciate your cooperation and hope to achieve even better results together!
Irina S. Druzhinina, Nanjing, China
Benefit from the IKOSA software solutions now and accelerate your microbiology research project
Developed especially for the needs of life science research institutions, the IKOSA Platform represents a complete software solution for the analysis of fungal biology images. The IKOSA Platform offers a broad range of tools suited for application in laboratory, industrial and educational settings. This one-of-a-kind product provides a wide spectrum of tool features including advanced viewing and analytical options, smart data storage and management and team collaboration. Test the IKOSA Platform now and accelerate your microbiology research project. Optimize the analysis process of fungal image data and get accurate results in no time.
We would like to thank Prof. Irina Druzhinina, and the following project team Fungal Genomics Group, at the Nanjing Agricultural University for providing images and graphs and the opportunity to use them in this interview.