PhenoComp

PhenoComp identifies population-level differentially expressed genes (DEGs) and infers their dysregulation directions from one-phenotype gene expression datasets lacking normal control samples.

Key Features:

• Population-level DEG detection: Identifies DEGs across a cohort using only one-phenotype datasets without requiring normal control samples.
• RankComp-based optimization: Extends the RankComp algorithm from individual-level DEG identification to population-level analysis.
• Direction inference: Determines dysregulation directions (up- or down-regulation) of DEGs across a disease-specific cohort.
• Comparative benchmarking: Shows performance comparable to significance analysis of microarrays, edgeR, and limma when case-control samples are used as the gold standard.
• Measurement-platform independence: Performance is reported to be consistent regardless of the measurement platform employed.
• Sensitivity to weak signals: Detects weakly differential expression signals in simulated and real datasets.
• Applicability to tissues without controls: Suitable for analysis of tissues where obtaining normal samples is difficult, such as heart and brain.

Scientific Applications:

• One-phenotype DEG discovery: Identification of DEGs in studies that lack matched normal control samples.
• Cohort-level dysregulation analysis: Inferring up/down regulation patterns across disease-specific cohorts.
• Tissue-specific studies without controls: Analysis of gene expression in tissues such as heart and brain where normal samples are scarce.
• Detection of subtle expression changes: Investigations requiring sensitivity to weak differential expression signals.

Methodology:

Optimizes the RankComp algorithm to detect population-level DEGs and infer dysregulation directions, and evaluates performance on simulated and real one-phenotype datasets with comparisons to significance analysis of microarrays, edgeR, and limma using case-control samples as the gold standard.