Artificial intelligence-enabled multi-scale virtual cell: perspective, challenges, and opportunities

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Traditional biological research faces limitations in spatiotemporal resolution and processing power for understanding dynamic cross-scale biological events.
• Artificial intelligence (AI) advancements have led to the development of the AI virtual cell (AIVC) as a novel research approach.
• AIVC integrates multi-omics data and multidisciplinary models to create digital twin systems for simulating cell functions and behaviors.

Purpose of the Study:

• To propose a unified definition and technical framework for AI virtual cells (AIVC).
• To analyze the cross-scale coupling mechanisms within the "gene-protein-pathway-cell" hierarchy.
• To provide a comprehensive overview of existing AIVC models and datasets.

Main Methods:

• Decomposition of the AIVC technical construction framework into cross-scale representation engineering, functional submodule design, and multi-component dynamic regulation.
• Integration of multi-omics data and multidisciplinary models.
• Analysis of cross-scale coupling mechanisms in biological systems.

Main Results:

• A proposed unified definition and technical framework for AIVC.
• Detailed analysis of "gene-protein-pathway-cell" hierarchy cross-scale coupling mechanisms.
• Summary of current AIVC models and datasets, identifying key challenges like data heterogeneity and model interpretability.

Conclusions:

• AIVC represents a significant shift in life sciences, moving towards predictability and innovation.
• Addressing AIVC's challenges is crucial for accelerating research progress.
• This review provides a foundational framework to guide future AIVC development and application.