"I like professional challenges," admits Lucie Klímová, who represented FIT at the 8 from BUT

On Wednesday, December 3, 2025, another annual presentation of the best bachelor's theses by students of Brno University of Technology 8 from BUT took place at the BUT Rector's Office. And this year, our faculty also had its representative at the gala evening. Lucie Klímová impressed with her work in the demanding field of bioinformatics. This traditional university event has two objectives: to recognize high-quality research work carried out during bachelor's studies and to offer successful students the opportunity to improve their presentation skills.

The theses and their authors are nominated by the faculty management. Selected students first have the opportunity to participate in a presentation skills course. During the gala evening, they then present their final theses to the audience – at this point, the main focus is on their ability to present the topic they have been working on in a clear and engaging way. The audience then votes for the Audience Award, which this year went to Denisa Zezulová from the Faculty of Business and Management at Brno University of Technology for her presentation on "Sales support and advertising for a specific company."

Don't be afraid of interdisciplinarity

Lucie Klímová presented her bachelor's thesis "Automated Techniques in DNA Analysis," supervised by Doc. Lukáš Holík. According to the author, she likes professional challenges, and this thesis definitely meets that criterion. Bioinformatics attracts few students due to its demanding nature, but Lucie wanted to pursue this field even before she began her studies at FIT BUT: "I enjoyed biology in high school and found interdisciplinarity to be a challenge." At the faculty, she seized the opportunity and did not hesitate when an interesting project assignment in this field appeared in the information system. She then further developed the topic in her bachelor's thesis.

Lucie Klímová's work is based on the application of finite automata to LTR retrotransposon search procedures. A little genetics theory is in order at this point: Transposons are repetitive DNA sequences in the genome. They are characterized by their ability to change position in the sequence, which is why they are often referred to as "jumping genes." They are a very old part of the genetic makeup of organisms. Studying them provides insight into the evolution of the genome. Transposons are highly represented in the genome (up to 40% of the human DNA sequence). At first glance, they do not appear to have a key function, and the term "junk DNA" was sometimes used in the past, but today we know that most of these sequences do have a function (albeit not a structural one), and it is precisely through study that we can better understand them. Their ability to translocate (change position in the sequence) influences the plasticity of the genome and its evolution in general. In addition, their position can influence adjacent parts of the sequence, including those that carry genetic information. The detection of LTR transposons can thus aid in the research of specific sequences and is a method used in modern genetics.

{{foto|VUT_RE_8zVUT_2025-12-03_Konicek_003.JPG|800}}

"We started with the TE-greedy-nester tool, which is used to search for transposons, and I identified a sub-algorithm in it that took up the most time when running the program, roughly 80% of the process. And we decided to rethink it with the aim of saving a significant amount of time," says Lucie, defining the intention of her bachelor's thesis in the most general terms. The aforementioned "sub-part" was performed by BlastX, perhaps the most widely used bioinformatics tool for searching gene sequences. The basic principle on which Lucie based her research was the idea that genome sequences should generally be representable by a finite automaton. Using an automaton for searching should then be significantly faster. However, the process of creating the automaton itself was challenging, the author recalls: "Transposon sequences often contain mutations. Finite automata are better for searching for exact matches, not similar matches, which is precisely what mutations complicate." Lucie based her work on another software tool frequently used in bioinformatics, HMMER, which uses hidden Markov models. These can be simplified as more complex versions of finite automata working with probabilities, in our case with mutation probabilities. "I then took the learned hidden Markov model and simplified it into a finite automaton in order to retain important information and speed up the search." Simplification was a challenge because it could not lead to the loss of important information about the nature of the sequence being searched for. Another problem was the non-determinism of the output, which corresponded to the nature of the probabilistic model. Lucie therefore had to deterministicize the resulting model, which, however, led to an output in the form of a giant model. And that, in her words, was a critical point that took a long time to deal with.

Challenges remain

The main result of Lucie Klímová's work is therefore the design of an algorithm for creating the aforementioned finite automaton. Among her research achievements, the author can credit the fact that the resulting finite automaton allows searching for transposon structural domains up to ten times faster than the BlastX tool. Once integrated into the TE-greedy-nester tool, this could lead to a significant acceleration in the search for LTR transposons. Lucie herself modestly points out that there is still much room for further improvement of the algorithm. For example, repeated sequence browsing offers the possibility of parallelization and further acceleration. It would also be appropriate to add result quality metrics that answer the question of how far the identified sequence is from the searched sequence. The author herself confirms that she would like to continue her research in the field of bioinformatics. She also mentions doctoral studies.

We will continue to keep our fingers crossed for our representative in the 8 from VUT competition on her professional path. And, of course, we thank her for representing our faculty so well. By the way: Lucie does not know who nominated her for the 8 from BUTcompetition. The person in question has, of course, the opportunity to come forward.

Let us remind you that FIT BUTalso left a significant mark in last year's competition, when Sára Jobranová took first place.