ProtRank

ProtRank performs differential abundance analysis by ranking proteins in proteomic and phosphoproteomic datasets while directly incorporating missing values to avoid imputation and infinite fold-change artifacts.

Key Features:

• Implementation: Implemented as a Python package for analysis of proteomic and phosphoproteomic data.
• Missing-value handling: Directly incorporates missing values into the analysis framework, avoiding imputation.
• Ranking-based differential analysis: Ranks proteins by their observed changes relative to other proteins to identify differential expression.
• Avoids infinite fold-change artifacts: Bypasses imputation that can produce artificial infinite fold-change calculations.
• Benchmarking: Produces results reported to be comparable to edgeR in evaluations.
• Evaluation: Validated on two distinct datasets to assess robustness to missing values.

Scientific Applications:

• Differential abundance analysis: Performs differential expression/abundance analysis for proteomic and phosphoproteomic datasets.
• Protein-level discovery: Identifies differentially expressed proteins without imputing missing data.
• Method comparison: Supports comparative benchmarking of differential analysis methods through comparison to edgeR.
• Missing-value scenarios: Applicable to analyses where missing values are present and imputation would alter downstream results.

Methodology:

Implemented in Python; ranks proteins based on observed changes relative to other proteins and directly incorporates missing values into the analysis framework; evaluated on two datasets and compared to edgeR.