ARTEMIS: A New Tool Unlocks Hidden RNA Structure Patterns
Researchers from the Laboratory of Bioinformatics and Protein Engineering at IIMCB have developed ARTEMIS - a new computational method with the potential to significantly advance our understanding of RNA, a key molecule involved in various cellular functions. 
RNA is not merely a simple chain of nucleotides that carries genetic information. RNA molecules also fulfill other biological functions, often requiring the formation of complex three-dimensional structures, much like proteins do. However, comparing these intricate shapes has proven challenging, especially when the molecules differ in nucleotide sequences. Recent advances in structural biology now provide scientists with tools to study RNA 3D structures more effectively and to better understand how these structures enable RNA to function within living organisms. If two RNA molecules exhibit similar structures, a hypothesis can be proposed and tested that they may also share similar functions. This is where ARTEMIS comes in.
From ARTEM to ARTEMIS
Dr. Eugene F. Baulin, who led the project in Prof. Janusz Bujnicki’s laboratory, recalls how he was inspired to create ARTEMIS. In 2022, the team designed a new algorithm called ARTEM, based on the idea that in the best possible spatial overlay of two RNA structures, at least one nucleotide from each structure would be nearly ideally superimposed on its counterpart. The research, described in a Nucleic Acids Research paper from 2023, states in the acknowledgments section: “Eugene F. Baulin thanks his daughter Leia for a walk in the park during which he came up with the idea of the ARTEM algorithm.”
Photo: Dr. Eugene F. Baulin, Laboratory of Bioinformatics and Protein Engineering
“It’s true,” Baulin confirms. “It happened in Szczęśliwicki Park in Warsaw while I was walking with my one-year-old daughter. Until that moment, I had occasionally thought about this research problem on and off for almost two years.”
Since before his PhD defense, Baulin had been interested in RNA tertiary motifs – spatially similar fragments of RNA found in various classes of structured RNA molecules, which consistently display the same core features across different structural environments. “You can think of it as, for example, different organs in different organisms. When we identify a new species, we want to describe it by comparing it to other, previously known species. What kind of organs does it have? For example, does it have eyes? If so, how do they function? Do they work like human eyes, or do they have a different structure, processing visual signals differently so that the organism can see something we cannot? We can ask similar questions about RNA structures,” explains Baulin. “Once we determine a new 3D structure of a particular RNA, perhaps one with an unknown function, annotating its motifs can help us understand its organization. This can often suggest possible functions. For instance, if the new RNA features a motif known to bind a certain molecule in other structures, we can speculate that this RNA might bind that same or similar molecule.” 
Fig. (left) optimal superposition of two RNA structures identified by ARTEMIS; (right) optimal structure-based sequence alignment identified by ARTEMIS. One of the RNA sequences is stretched from left to right, another sequence is stretched from top to bottom, and each dot (square) depicts a pair of residues matching each other in the structure superposition.
The ARTEM algorithm was designed to identify local structural similarities. The next step was the development of ARTEMIS, to enable comparisons of entire RNA molecular structures and determine which nucleotides in one sequence correspond to nucleotides in another, creating so-called sequence alignment. Baulin and collaborating team members conducted preliminary tests and found that ARTEMIS works very well—and for many RNA molecules, even better than the best tools previously developed by other researchers. Through the comparison of RNA structures, ARTEMIS can accurately determine which nucleotides correspond to each other. This confirms that the research assumption of identifying structural motifs by aligning individual nucleotides and checking how the remaining nucleotides relate to each other yields excellent results in comparing the spatial structures of entire RNA molecules.
Peeking at RNA’s Secrets
ARTEMIS was specifically designed to analyze RNA’s three-dimensional shapes, even in cases where the sequences differ significantly, including changes in nucleotide order. Current tools often struggle with such a task. In contrast, ARTEMIS operates without restrictions on the order of nucleotides in the sequences being compared, allowing scientists to compare RNA structures more precisely and effectively.
The new method may interest researchers in structural bioinformatics, who analyze structural databases, classify known structures, and predict the structures of RNA molecules that have not yet been determined. It is also a valuable tool for structural biologists who experimentally determine new structures of nucleic acids, both RNA and DNA. These scientists can use ARTEMIS to search structural databases for similar structures or motifs that may provide insights into biological functions.
Artemis can also contribute to the improved design of new functional RNA molecules and molecules that interact with RNA, applicable in biotechnology and medicine. As Baulin notes, practically every introduction of a truly innovative invention into everyday use stems from earlier basic research - such as that which led to the development of the ARTEMIS method.
“What excites me most,” says Eugene F. Baulin, “is how one simple assumption about RNA structure led to the development of powerful tools that allow us to make new discoveries. It shows how a deep understanding of data can lead to a simple solution that outperforms more complex approaches. With the success of the AlphaFold artificial intelligence algorithm in predicting protein structures, RNA structure prediction has now become the next holy grail of computational biology, and each new tool, like ARTEMIS, brings us a little closer to that solution,” concludes the scientist from the Laboratory of Bioinformatics and Protein Engineering at the International Institute of Molecular and Cell Biology in Warsaw.
An article “ARTEMIS: a method for topology-independent superposition of RNA 3D structures and structure-based sequence alignment” in Nucleic Acids Research is available here: https://academic.oup.com/nar/article/52/18/10850/7754985