GANSamples-ac4C: Enhancing ac4C site prediction via generative adversarial networks and transfer learning

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• N4-acetylcytidine (ac4C) is a crucial RNA modification impacting cellular functions.
• Experimental detection of ac4C sites is costly and resource-intensive.
• Accurate computational prediction of ac4C sites is challenging due to limited training data.

Purpose of the Study:

• To develop a novel framework, GANSamples-ac4C, for enhancing ac4C modification site prediction.
• To leverage transfer learning and generative adversarial networks (GANs) to generate synthetic RNA sequences.
• To improve the accuracy and stability of ac4C prediction models by overcoming data scarcity.

Main Methods:

• Developed GANSamples-ac4C, a framework combining transfer learning and GANs.
• Generated synthetic RNA sequences to augment limited experimental data.
• Evaluated the framework's performance against existing state-of-the-art methods.

Main Results:

• GANSamples-ac4C significantly outperforms current methods in identifying ac4C sites.
• The study demonstrates the efficacy of synthetic data in addressing data scarcity for biological sequence prediction.
• Interpretability analyses identified key sequence regions and motifs influencing ac4C prediction.

Conclusions:

• Synthetic data generation using GANs is a viable strategy for improving RNA modification prediction models.
• GANSamples-ac4C offers a powerful and interpretable tool for ac4C site identification.
• The framework provides novel insights into ac4C modification mechanisms and associated sequence features.