Stacking-ac4C: an ensemble model using mixed features for identifying n4-acetylcytidine in mRNA.

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• N4-acetylcytidine (ac4C) is a crucial mRNA modification impacting translation.
• The exact mechanisms of ac4C modification and its role in translation are not fully understood.
• Current experimental methods for identifying ac4C sites are inefficient and time-consuming.

Purpose of the Study:

• To develop a rapid and accurate computational method for identifying ac4C sites in mRNA.
• To address the need for efficient tools in ac4C-related molecular biology research.

Main Methods:

• Development of a Stacking-based heterogeneous integrated ac4C model (ST-ac4C).
• Integration of three feature extraction techniques: Kmer, electron-ion interaction pseudo-potential values (PseEIIP), and pseudo-K-tuple nucleotide composition (PseKNC).
• Utilization of the Cluster Centroids algorithm to handle imbalanced data and mitigate underfitting.

Main Results:

• The ST-ac4C model demonstrated significant improvements in performance metrics, including a 15.61% increase in Matthews Correlation Coefficient (MCC) and a 5.97% increase in the Receiver Operating Characteristic (ROC) curve compared to existing models.
• On a balanced dataset, the model achieved a 4.1% increase in sensitivity (Sn) and a nearly 1% increase in accuracy (Acc), with a 0.35% ROC improvement.
• Independent testing confirmed the model's superior predictive capability for ac4C sites.

Conclusions:

• The proposed ST-ac4C model offers a robust and efficient solution for identifying ac4C sites.
• This computational tool can accelerate research into the functional roles of ac4C modifications in mRNA translation.
• The model's open accessibility facilitates its adoption and further development in the scientific community.