What it does

Performs end-to-end differential expression analysis for LFQ proteomics: input parsing for MaxQuant and DIA-NN outputs, filtering of contaminants, log2 transformation, down-shifted Gaussian imputation, statistical testing (two-sample t-test with s0-based FDR correction), and diagnostic visualizations (PCA, volcano, imputation diagnostics). Produces a reproducible report directory with figures and tables.

When to use it

Use this skill when you have processed proteinGroups (MaxQuant) or DIA-NN output and need a reproducible DE analysis pipeline (e.g., quick exploratory DE on treatment vs control contrasts). Not for raw MS processing or biological interpretation — it's an analysis/visualization pipeline that requires user review of results.

What's included

Scripts: examples and a proteomics_de.py demo script present in the repo
References: domain decisions and statistical references included in the skill body
Instructions: detailed preprocessing rules, imputation strategy (shift=1.8, scale=0.3), default thresholds (FDR=0.05, s0=0.1), and output structure for reports.