Accelerated knowledge discovery from omics data by optimal experimental design.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Designing informative experiments for complex biological systems is challenging.
• Accelerating knowledge discovery requires efficient data generation and analysis.

Purpose of the Study:

• To present an optimal experimental design (OPEX) method using machine learning.
• To guide omics data collection for improved predictive modeling and knowledge discovery.

Main Methods:

• Developed OPEX, an active learning approach for experimental design.
• Applied OPEX to explore Escherichia coli populations under biocide and antibiotic stress.
• Utilized machine learning for experimental space exploration and model training.

Main Results:

• OPEX achieved 44% data reduction for accurate gene expression predictive models.
• Identified broad exploration followed by fine-tuning as an optimal experimental strategy.
• Discovered 29 instances of cross-stress protection and 4 of cross-stress vulnerability in E. coli.

Conclusions:

• Active learning can guide omics data collection for evidence-driven decisions.
• OPEX accelerates knowledge discovery in life sciences by optimizing experimental design.
• Identified key roles for chaperones, stress proteins, and transport pumps in cross-stress responses.