PseU-KeMRF: A Novel Method for Identifying RNA Pseudouridine Sites

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Pseudouridine is a prevalent RNA modification essential for diverse biological functions.
• Accurate identification of pseudouridine sites is critical for understanding its biological roles.
• Traditional experimental methods for pseudouridine site identification are challenging.

Purpose of the Study:

• To develop a fast and accurate computational method for predicting pseudouridine sites from RNA sequences.
• To introduce the novel PseU-KeMRF model for pseudouridine site prediction.
• To evaluate the model's performance across multiple species.

Main Methods:

• Utilized four RNA coding schemes: binary feature, position-specific trinucleotide propensity based on single strand (PSTNPss), nucleotide chemical property (NCP), and pseudo k-tuple composition (PseKNC).
• Employed support vector machine-recursive feature elimination (SVM-RFE) for feature selection and optimization.
• Applied the kernel based on multinomial random forests (KeMRF) classifier for prediction and validation.

Main Results:

• The PseU-KeMRF model demonstrated superior predictive performance compared to existing models.
• The KeMRF classifier improved upon traditional random forests through enhanced node splitting and kernel integration.
• The model successfully identified pseudouridine sites in *H. sapiens*, *S. cerevisiae*, and *M. musculus*.

Conclusions:

• PseU-KeMRF is a highly competitive and effective computational tool for identifying pseudouridine sites in RNA sequences.
• The developed model facilitates further research into the biological mechanisms of pseudouridine.
• This computational approach addresses the limitations of traditional experimental methods.