Essential genes identification model based on sequence feature map and graph convolutional neural network.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Essential genes are critical for organismal life, including growth, development, and cellular functions.
• Traditional methods for identifying essential genes are laborious and time-consuming.
• Current machine learning models for essential gene prediction require improved accuracy.

Purpose of the Study:

• To develop a robust computational model for accurate essential gene prediction.
• To enhance the identification of essential genes using advanced machine learning techniques.

Main Methods:

• Developed GCNN-SFM, a model utilizing graph convolutional neural networks (GCNN).
• Integrated graph convolutional, convolutional, and fully connected layers for feature extraction from gene sequences.
• Employed coding techniques to transform gene sequences into feature maps for GCN analysis.
• Utilized gradient descent to optimize the cross-entropy loss function for enhanced prediction accuracy.

Main Results:

• GCNN-SFM demonstrated superior performance compared to existing advanced essential gene prediction models.
• Achieved an average prediction accuracy of 94.53% in experimental evaluations.
• Effectively captured both local and global features of gene sequences through multi-layer GCN operations.

Conclusions:

• GCNN-SFM offers a novel and effective computational approach for essential gene identification.
• The model's high accuracy has significant implications for advancing biology and genomics research.
• This work provides a valuable tool for researchers studying gene essentiality.