SPICE: discovery of phenotype-determining component interplays.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Cellular behavior is complex and understanding its fundamental rules is a major goal in systems biology.
• Identifying system components and their interactions that determine phenotypes is a challenging but crucial step.

Purpose of the Study:

• To develop a data-driven method for enumerating phenotype-determining system component interplays.
• To improve the prediction of biological system phenotypes by identifying key contributing components.

Main Methods:

• Proposed the System Phenotype-related Interplaying Components Enumerator (SPICE) approach.
• SPICE iteratively identifies statistically significant system components crucial for phenotype specificity, functional coherence, and predictive accuracy.
• The method is applicable to both instance-based and network-based data.

Main Results:

• SPICE successfully identified system components related to biohydrogen production, motility, and cancer phenotypes.
• Manual curation confirmed known phenotype-related components reported in scientific literature.
• Incorporating identified components as features improved phenotype classification accuracy by 10% compared to state-of-the-art methods on benchmark microarray datasets.

Conclusions:

• SPICE provides an effective methodology for enumerating phenotype-determining system component interplays.
• The method demonstrated improved identification of phenotype-associated gene sets for cancer, biohydrogen production, and motility.
• SPICE enhanced predictive accuracy for system phenotypes compared to various individual classifiers and ensemble methods.