rpiCOOL: A tool for In Silico RNA-protein interaction detection using random forest.

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• RNA-protein interactions (RPIs) are crucial for post-transcriptional gene regulation and understanding complex diseases.
• Experimental determination of RPIs is challenging, necessitating computational prediction methods.

Purpose of the Study:

• To develop a novel machine learning method for predicting RPIs based on sequence information.
• To identify informative sequence-derived features for accurate RPI prediction.

Main Methods:

• A random forest classifier was employed using sequence composition, motif information, and repetitive patterns as descriptors.
• Feature selection methods identified informative features, with sequence motifs and nucleotide composition proving effective.
• The proposed method, rpiCOOL, was evaluated using 10-fold cross-validation on benchmark datasets.

Main Results:

• The rpiCOOL method demonstrated superior performance compared to existing state-of-the-art methods across various metrics.
• Sequence motifs and nucleotide composition were identified as significant contributors to RPI prediction accuracy.
• Prediction accuracy varied significantly across different organisms, highlighting the need for organism-specific models.

Conclusions:

• The developed machine learning approach effectively predicts RNA-protein interactions using sequence-based features.
• rpiCOOL provides a valuable, user-friendly tool for researchers to predict RPIs, advancing gene regulation and disease studies.
• The findings underscore the importance of sequence features and organism context in RPI prediction.