Machine learning methods for gene regulatory network inference.

Area of Science:

• Computational Biology and Bioinformatics
• Genomics and Systems Biology

Background:

• Gene Regulatory Networks (GRNs) govern gene expression in response to biological signals.
• High-throughput sequencing and computational biology have advanced GRN inference.
• Artificial intelligence (AI), especially machine learning (ML), is crucial for analyzing omics data to understand gene interactions.

Purpose of the Study:

• To provide a comprehensive review of ML-based GRN inference methodologies.
• To support the application of GRN inference and the development of new ML methods.
• To discuss datasets, evaluation metrics, and challenges in GRN inference.

Main Methods:

• Review of supervised, unsupervised, semi-supervised, and contrastive learning techniques for GRN inference.
• Analysis of commonly used datasets and evaluation metrics in the field.
• Emphasis on deep learning approaches for enhanced inference performance.

Main Results:

• AI and ML techniques significantly improve the accuracy of GRN inference.
• Deep learning methods show promise for enhancing the performance of GRN inference.
• Identified common datasets and metrics for evaluating GRN inference models.

Conclusions:

• ML, particularly deep learning, is a powerful tool for deciphering complex GRNs.
• Further research into novel ML methods and addressing current challenges will advance GRN inference.
• This review serves as a guide for researchers in GRN inference and ML development.