UPWr doctoral students awarded research grants

• Aleksandra Kot will study the changes taking place in the soil in Iceland as a result of human activity and climate change
• Agnieszka Krawczyk-Łebek will check whether the chlorine atom enhances the antimicrobial activity of flavonoid glycosides
• Karolina Tkacz will investigate the possibilities of reducing postprandial hyperglycemia with bioflavonoids
• Kamil Smolak will develop an algorithm to study human mobility
• Agata Walicka will develop methodology essential for the analysis of data obtained from airborne laser scanning

In the Preludium 20 competition organised by the National Science Centre, eleven doctoral students of the Wrocław University of Environmental and Life Sciences received nearly PLN two million for their research. Below we present another 5 of them.

Iceland instead of the Karkonosze Mountains

Aleksandra Kot from the Institute of Soil Science, Plant Nutrition and Environmental Protection is conducting research in Iceland.

– After defending my master's thesis, during my "gap year", when I was wondering whether to pursue a PhD, I went to Iceland for a few months. I worked at a hostel there. Its owner, Hannes, turned out to be a man of great knowledge about local nature and passionate about science. He was eager to share it with me, and showed me some interesting places in the area, indicating the changes taking place in this region. It was then that I came up with the idea of research on Iceland's natural environment, which was enthusiastically accepted by my thesis supervisor at that time, Dr Jarosław Waroszewski – says Aleksandra Kot with a smile, adding that not only will a research project from the Prelude 20 programme be carried out in Iceland, but also another one from the Opus 20 competition, which is headed by Dr Waroszewski, the supervisor of her doctoral thesis.

– And these two projects would not exist if it were not for Hannes, because my PhD was originally supposed to be about research on soils of the Karkonosze Mountains, while I am analysing changes in the soils of south-eastern Iceland under the influence of both climate change and direct human activity in this area, mainly related to cattle and sheep grazing, and hay production – explains the doctoral student from the Institute of Soil Science, Plant Nutrition and Environmental Protection.

Subarctic and arctic areas are extremely vulnerable to ongoing climate changes. In addition, the volcanic soils that occur in Iceland have high levels of organic matter, constituting a "storage" of organic carbon. The project involves two types of interrelated research. One of them will concern WEOM, i.e. water-soluble organic matter, which is the most dynamic and bioavailable fraction of organic matter, responsible for the activation and movement of many elements in the soil. WEOM influences the sequestration of  C02 from the atmosphere and the release of greenhouse gases from the soil. Regular gas emission measurements are the second type of test. One of the hypotheses of the project is the assumption that as a result of the ongoing processes, volcanic soils cease to function as carbon storage, and even become its source.

– This organic matter is influenced by temperature, i.e. climate change, but also by usage related, among others, to soil fertilisation. Research is also related to obtaining information from local residents. We want to find out how the researched area was used, i.e. whether animals were grazed there, hay was obtained, whether fertilisers were used, and, if yes, what kind of fertilisers – says Aleksandra Kot, adding that the research points will be located near the largest glacier in Europe – Vatnajökull, so that they reflect diversification of use and mosaic of natural habitats. The collected soil samples will be analysed for WEOM, but the microbiological biomass of carbon and nitrogen will also be checked. Aim of the study: characteristics of environmental changes in terms of their further progression, as well as the effects of global warming in polar and subpolar regions. The research of the Polish doctoral student is of interest to scientists from Iceland, who are willing to cooperate in the project.

Chlorine versus microbes

Agnieszka Krawczyk-Łebek from the Department of Chemistry, whose research is conducted under the scientific supervision of Prof. Edyta Kostrzewa-Susłow, concerns the chemical and microbiological synthesis of flavonoid glycosides with the chlorine atom.

– Flavonoids with a chlorine atom are rarely found in nature, and their biological activity has not yet been fully investigated. One of the few examples of such compounds is the highly antimicrobial chloroflavonin produced by filamentous fungi of species such as Mucor irregularis and Aspergillus candidus. Hence the idea of research on the production and biological activity of this group of compounds – says Agnieszka Krawczyk-Łebek, who first obtains flavonoids with a chlorine atom through chemical synthesis, and then subjects them to microbial glycosylation with entomopathogenic filamentous fungi in order to improve their solubility in water and bioavailability.

– The next stage of the research will be preliminary tests of the antimicrobial activity of the obtained flavonoids against selected strains of pathogenic bacteria of the species Escherichia coli and Staphyloccocus aureus and yeasts of the species Candida albicans, as well as against probiotic bacteria beneficially affecting the digestive system, which will enable assessing whether the tested compounds are not only effective but also safe. In the summer of 2022 I will also conduct biological research on the obtained compounds in cooperation with the University of Porto, as part of the Erasmus+ programme. The search for new, safe and effective therapeutic agents is extremely important in the context of the increasing drug resistance of many pathogenic microorganisms, and flavonoids with a chlorine atom are a promising subject of research in this field – emphasises Agnieszka Krawczyk-Łebek.

Polyphenols for hyperglycemia

Karolina Tkacz, who works in the Plants4FOOD team under the supervision of Prof. Aneta Wojdyło and Prof. Paulina Nowicka, will in turn deal with postprandial hyperglycemia, and more specifically the possibilities of its effective reduction through the active action of polyphenolic compounds in synbiotic formulas.

– On the one hand, I was interested in this research area, because postprandial hyperglycemia, and thus the metabolic syndrome and diabetes, is a serious problem not only for our health, but also the economy. After all, it is said that we live in times of an epidemic of civilisation diseases. On the other hand, in the course of research in the Plants4FOOD team, we proved that polyphenolic compounds have antidiabetic potential, and are distinguished by high activity towards inhibiting enzymes involved in the breakdown of polysaccharides – says Karolina Tkacz, whose doctoral studies went off to a flying start when she won the Diamond Grant.

As the doctoral student from the Department of Fruit, Vegetable and Plant Nutraceutical Technology explains, the research also shows that the human body absorbs only 20 percent of polyphenolic compounds from food.

– This poor level of bioavailability is both a problem and a challenge, so I will work on increasing it, based on the most recent synbiotic concept, referring to a mixture of living microorganisms and ingredients that provide health benefits by using the body's microflora. Polyphenolic compounds fit perfectly into the second part of the tandem, i.e. the prebiotic factor. So far, however, research on the prebiotic properties of polyphenols has been carried out with one area in mind: testing their effect on bacterial survival and microbiota modulation – says Karolina Tkacz, who in her project has decided to do everything the other way around, i.e. she will check how selected probiotic strains affect the polyphenolic compounds contained in plant extracts.

– Enzymatic modifications via bacteria, digestive conditions and probiotic-prebiotic interactions will promote the formation of phenolic metabolites of interest to us, with higher bioavailability and new pro-health activity – adds Karolina Tkacz, emphasising that synbiotics show potential as therapeutic agents in reducing postprandial hyperglycaemia occurring in diabetes, affecting as much as 9 percent of the population. Research to date has focused on conventional prebiotic agents, oligosaccharides and inulin, the use of which is gradually being restricted – it has been found that they promote the growth of pathogenic microorganisms and have an ambiguous effect on the human body. Hence the need to look for new solutions, and the list of potentially beneficial compounds includes polyphenols, which need to have their bioavailability and degree of absorption enhanced.

Mobility algorithm

Explaining and mitigating statistical bias in human mobility forecasts – in his research work, Kamil Smolak from the Institute of Geodesy and Geoinformatics analyses data from mobile devices such as mobile phones or sports bands, pointing out that he does not do this on an individual level, but collectively.

– While browsing literature, I noticed that the same algorithms that are designed to predict where people will be in an hour or two, when applied to different data lead to totally different results. It turned out that how the data is processed impacts the results obtained, which means that they can be easily manipulated. On the one hand, intelligently prepared data allows algorithms to achieve up to 99 percent accuracy, but poorly prepared data causes statistical load on the data, which means that these results are unreliable. As a result, they cannot be compared – says Kamil Smolak, explaining that this phenomenon is scientifically called the problem of a variable spatial reference unit, and was applied in practice by the American governor Elbridge Gerry, who in the 19th century manipulated constituencies in such a way that in each of them his party obtained the majority of votes.

– The problem of a variable spatial reference unit concerns such manipulation of data that enables achieving an anticipated result. It concerns the spatial dimension, but also the time dimension. This is what we observe in research on human mobility – says the doctoral student from the Institute of Surveying and Geoinformatics, who initially will analyse the phenomenon itself, but in the second stage will develop a method that, regardless of the spatial and temporal reference, will explain how the trajectory of motion and its changes will change parameters, such as its predictability. The obtained method is to be used as a basis for the development of an algorithm that will search for optimal parameters for mobile data processing, allowing for the preservation of a maximum amount of information – this is the third stage of the project.

– The results of the algorithms on the prepared data are very high, because information is removed from them. I want to keep as many of them as possible, because this will increase the usefulness of the data for many analyses, such as urban spatial planning, location of road networks, or modelling the spread of infectious diseases. And these are just a few of the areas in which the analysis of people movement, traffic direction, intensity and time is important – emphasises Kamil Smolak, whose supervisor is Prof. Witold Rohm.

How to check the point cloud

In her project, Agata Walicka will develop a methodology based on deep learning, allowing for repetitive classification of airborne laser scanning point clouds with different characteristics. The supervisor of her doctorate was the late Prof. Andrzej Borkowski, and her current supervisors are Prof. Grzegorz Jóźków and Prof. Norbert Pfeifer.

– We perceive the surrounding reality as a set of objects located in three-dimensional space. The characteristics of the geometric properties of these objects and their location in space require the determination of their dimensions and coordinates that indicate their position. Remote sensing can be used for this. The results of measurements can be presented in various ways, but the most similar to a 3D perception of the world by a human being is the so-called point cloud – explains Agata Walicka, adding that this cloud is an unordered set of points representing the surface of measured objects, where each point is assigned three coordinates that describe its location in 3D space.

This data is collected in various ways. In her research, Agata Walicka will deal with airborne laser scanning, which is used to measure large objects, such as entire cities.

– And this is where a problem arises, because the point cloud obtained in this way shows the surface of a very large number of objects of various types, such as buildings, trees, cars or street lamps. In order to be able to further process this data, it is necessary to perform another step in the analysis – to distinguish, for example, buildings from trees, and therefore assign to each point a description indicating its belonging to a certain group of objects, so-called classes – vegetation, land area, buildings, etc. This step is called classification. Humans do this instinctively – it is enough to look at the visualisation of the point cloud and recognise particular objects. However, this is much more complicated for a computer. Therefore, in order to perform automatic classification, it is necessary to manually prepare so-called learning data, i.e. an example of the correct assignment of points to individual classes for some of the obtained data. The computer is then 'trained' to correctly assign points to specific classes based on the learning data. However, manual preparation of the training data is tedious and time-consuming. Therefore, in my project I want to develop methodology to minimise the need for preparing training data before performing each classification – explains Agata Walicka, who will use deep neural networks, i.e. artificial intelligence.

As the doctoral student from the Institute of Surveying and Geoinformatics emphasises, the original concept of neural networks was based on attempts to recreate the way the human brain works, and deep neural networks consist of many layers of individual neurons.

– At the same time, due to its complex nature, training a deep neural network is time-consuming and requires a large amount of training data. The development of such methodology that allows for one-off training of a neural network, and then using it to classify data obtained with different equipment and having different characteristics, would therefore eliminate these drawbacks – says Agata Walicka, who will develop the new methodology using publicly available data from airborne laser scanning.