DiCleavePlus: A Transformer-Based Model to Detect Human Dicer Cleavage Sites Within Cleavage Patterns.

Area of Science:

• Molecular Biology
• Bioinformatics
• Genetics

Background:

• MicroRNAs (miRNAs) are key regulators of gene expression at the posttranscriptional level.
• Mature miRNA biogenesis depends on precise cleavage of precursor miRNAs (pre-miRNAs) by the Dicer enzyme.
• Current computational methods for predicting Dicer cleavage sites have limitations, including suboptimal performance or reliance on restricted sequence patterns.

Purpose of the Study:

• To develop a more accurate computational framework for predicting human Dicer cleavage sites on pre-miRNAs.
• To overcome the limitations of existing Cleavage Pattern-based and non-pattern-based prediction models.
• To leverage both sequence and structural information of pre-miRNAs for improved prediction accuracy.

Main Methods:

• Introduction of DiCleavePlus, a novel Cleavage Pattern-based prediction framework.
• Inputting an extended Cleavage Pattern and the full-length pre-miRNA sequence into the model.
• Utilizing a Transformer-based encoder to extract relevant features from both the pattern and the pre-miRNA sequence.

Main Results:

• DiCleavePlus demonstrates accurate and robust performance in predicting human Dicer cleavage sites.
• The framework effectively integrates sequence and structural features for enhanced prediction.
• Benchmarking experiments confirm the superior performance of DiCleavePlus compared to existing approaches.

Conclusions:

• DiCleavePlus represents a significant advancement in predicting Dicer cleavage sites on pre-miRNAs.
• The model's ability to exploit extended sequence patterns and structural information offers improved accuracy.
• This tool has the potential to enhance research in miRNA biogenesis and gene regulation analysis.