Deep Learning in Gene Regulatory Network Inference: A Survey.

Area of Science:

• Genomics and Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Gene regulatory networks (GRNs) are essential for understanding biological processes like development and cellular response.
• Inferring GRNs from observational data is a key challenge in biological applications.
• The increasing volume and complexity of biological data necessitate advanced computational methods for accurate GRN inference.

Purpose of the Study:

• To provide a comprehensive survey and analysis of deep learning-based methods for gene regulatory network inference.
• To categorize existing deep learning GRN inference methods based on data applicability.
• To evaluate the effectiveness and scalability of these methods across various scenarios.

Main Methods:

• Systematic review and analysis of 12 prominent deep learning models for GRN inference.
• Categorization of methods based on the types of biological data they utilize (e.g., transcriptomic, epigenomic).
• Detailed examination of the core concepts, algorithms, and implementation steps for each method.

Main Results:

• Identification and classification of 12 key deep learning approaches for GRN inference.
• Comparative evaluation of method performance, highlighting strengths and limitations in different biological contexts.
• Assessment of scalability and computational efficiency across diverse datasets.

Conclusions:

• Deep learning offers powerful tools for advancing gene regulatory network inference.
• Method selection depends critically on data type and specific research questions.
• Future research should focus on addressing current challenges and developing more robust and scalable deep learning models for GRN inference.