limpa

limpa quantifies and performs differential analysis of mass-spectrometry proteomics data with probabilistic recovery of information from missing values, avoiding the need for imputation.

Key Features:

• Detection Probability Curve Estimation: limpa estimates the detection probability curve (DPC), which relates the probability of successful detection to the underlying log-intensity of each precursor ion.
• Missing Value Integration: The DPC is used to incorporate missing values into protein quantification and into subsequent differential expression analyses without imputation.
• Variance Modeling and Precision Weights: The uncertainty of protein quantifications is propagated through to limma analyses using variance modeling and precision weights, ensuring accurate error rate control.
• Flexible Input Data Types: The analysis pipeline accepts precursor or peptide intensities including missing values and can alternatively work with PTM or protein level data.
• Downstream limma Compatibility: limpa produces objects suitable for downstream analysis in limma, allowing complex experimental designs and other downstream tasks such as gene ontology or pathway analysis.
• Peptide-Level Summarization: limpa summarises peptide-level data to estimate an expression value for every protein in every sample.
• Precision-Based Downweighting: Expression values supported by fewer detected peptides or involving more missing values are treated as less precise and are downweighted in differential expression analysis.

Scientific Applications:

• Protein Quantification: limpa produces complete protein quantifications from precursor or peptide intensities containing missing values.
• Differential Expression Analysis: limpa enables differential expression analysis of MS proteomics data with accurate error rate control through variance modeling and precision weights.
• Single-Cell and Spatial Proteomics: limpa addresses the high frequency of missing values in emerging fields such as single-cell and spatial proteomics.
• Gene Ontology and Pathway Analysis: limpa enables downstream tasks such as gene ontology or pathway analysis through its production of limma-compatible linear model objects.

Methodology:

limpa estimates a detection probability curve (DPC) relating detection probability to underlying log-intensity of each precursor ion, uses this to incorporate missing values into protein quantification and differential expression analysis, summarises peptide-level data to protein-level expression values, applies variance modeling and precision weights to propagate quantification uncertainty, and produces linear model objects for downstream analysis with limma.