KR-20260062609-A - METHOD FOR PREDICTING RNA STRUCTURE

Abstract

A method for predicting RNA structure is disclosed, comprising: (a) defining an atomic system having one or more rotation angles obtained according to a predetermined RNA sequence; and (b) calculating the free energy of the system after adjusting one or more of the rotation angles in the atomic system, and repeating step (b) with an AI model to derive the atomic system with the lowest free energy.

Inventors

• 김주현

Assignees

• 김주현

Dates

Publication Date

20260507

Application Date

20241029

Claims (1)

1. (a) defining an atomic system having one or more rotation angles obtained according to a predetermined RNA sequence; and (b) a step of calculating the free energy of the system after adjusting one or more of the rotation angles in the atomic system; comprising, Repeating step (b) above with an AI model to derive the atomic system with the lowest free energy, Method for predicting RNA structure.

Description

Method for Predicting RNA Structure This is about a method for accurately predicting RNA structure. Proteins have become the focus of extensive research, leading to the development of AI capable of predicting various forms of protein folding. Furthermore, in-depth studies have been conducted on individual proteins. RNA plays a crucial role in transferring genetic information from DNA to proteins. However, due to relatively lower interest compared to proteins, ribonucleic acid (RNA) has been neglected in research, and consequently, no AI capable of accurately predicting RNA folding exists. Furthermore, because RNA exhibits significant structural variability, current AI models struggle to capture the fundamental principles of its three-dimensional structure. The best current AI models for predicting RNA structure show an accuracy of 40% to 80%. Meanwhile, since the structure of RNA consists of five major bases, which is simpler than that of proteins composed of 20 amino acids, manual programs based on data and experimental testing may outperform standalone AI models. The present invention was developed to solve these problems and predict RNA structure more accurately. FIG. 1 shows the twist angle of an RNA structure defined according to the prior art; FIG. 2 illustrates an example having a different twist angle according to one embodiment of the present specification; FIG. 3 is an exemplary illustration of another free energy having another form according to one embodiment of the present specification. Hereinafter, one aspect of the present specification will be described with reference to the attached drawings. However, the details described in the present specification may be implemented in various different forms and are therefore not limited to the embodiments described herein. Furthermore, in order to clearly explain one aspect of the present specification in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification have been given similar reference numerals. Throughout the specification, when it is stated that a part is "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "indirectly connected" with other members interposed between them. Furthermore, when it is stated that a part "includes" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but rather allows for the inclusion of additional components. When a range of numerical values is described in this specification, unless a specific range is otherwise described, the value has the precision of significant figures provided according to the standard rules in chemistry for significant figures. For example, 10 includes a range of 5.0 to 14.9, and the number 10.0 includes a range of 9.50 to 10.49. Hereinafter, one aspect of the present specification will be described in detail with reference to the attached drawings. Methods for predicting RNA structure A method for predicting an RNA structure according to one embodiment of the present specification comprises: (a) defining an atomic system having one or more rotation angles obtained according to a predetermined RNA sequence; and (b) calculating the free energy of the system after adjusting one or more of the rotation angles in the atomic system, and can derive the atomic system with the lowest free energy by repeating step (b) with an AI model. This invention predicts RNA folding structures by treating all possible rotation angles in the RNA structure as variables and free energy as the output value. By inversely calculating this function, the rotation angles of individual RNA molecules can be identified with unprecedented accuracy. This hybrid approach ensures high accuracy by handling each RNA individually instead of universally applying a single AI model. RNA structure exhibits seven angles that can be represented from α to ζ (α, β, γ, δ, ε, η, ζ). It also includes the χ angle, which is the relative rotation of the base and the sugar. These angles determine the ribose-phosphate backbone structure. Figure 1 shows an example of a prior art that defines such a twist angle. While the loop structure of bases is relatively fixed, the rotation angles discussed above are variable, so a mathematical approach to them is necessary for predicting RNA structure. To express rotation angles, they can be represented as mathematical coordinates (points). An example of the transformation process is as follows: 1. Represent three atoms as coordinates. 2. Define the axis of rotation using the first and second atoms. 3. Relocate the three atomic systems to the absolute coordinate system. 4. Align the system so that the rotation axis becomes the z-axis. 5. Rotate the system according to the rotation angle. 6. Apply the inverse transformation sequentially to return the system to its original position. Steps 3 through 6 are essential when