DeepAc4C

DeepAc4C employs a convolutional neural network to identify N4-acetylcytidine (ac4C) modifications in mRNA and eukaryotic RNA to support investigation of their roles in human disease.

Key Features:

• Convolutional Neural Network Architecture: Uses a CNN model to capture complex sequence patterns for ac4C site prediction.
• Hybrid Feature Composition: Integrates physicochemical properties of nucleic acids with distributed representation features to characterize sequence contexts.
• Balanced Performance Metrics: Demonstrates superior and more balanced predictive performance, reducing false positives and negatives relative to prior predictors.
• Identification of Sequence Motifs: Detects specific sequence motifs characteristic of ac4C modifications.

Scientific Applications:

• Disease Research: Enables molecular studies of ac4C associations with various human diseases.
• RNA Biology Studies: Supports investigation of functional roles of ac4C in gene expression regulation and epigenetics.

Methodology:

Processes input sequences with a convolutional neural network using hybrid features derived from physicochemical properties and distributed representations of nucleic acids to capture sequence patterns for ac4C prediction.