Research data for: The "Myasthenia Gravis Effect" in Thymoma - DIA serum proteomics (myasthenia gravis)
Description
DIA (data-independent acquisition) serum proteomics dataset supporting the manuscript "The Myasthenia Gravis Effect in Thymoma: A Serum Proteomic Panel of MG-Specific Proteins Revealed by DIA Mass Spectrometry" (Journal of Neuroimmunology). Study design: 157 serum samples across 10 clinically defined groups (712 proteins identified, 629 quantified). Main comparisons: thymoma with myasthenia gravis (MG) vs thymoma without MG (TMGOP vs TNMG, the "MG-effect layer"); thymic hyperplasia with MG vs healthy controls (HMGOP vs CON). Contents: differentially expressed protein (DEP) tables, per-protein statistics with rank-based AUC, panel performance (in-sample and leave-one-out cross-validation AUC), sensitivity analyses (pooled-sample exclusion), and Python analysis scripts for reproducibility. Key results: 46 DEPs (P<0.05, |log2FC|>=1) in the MG-effect layer; a 7-gene panel (ARF3, AZU1, FTL, KRT10, KRT2, KRT6A, KRT6B) shared across MG presentations with LOO-CV AUC 0.98-0.99. Raw mass-spectrometry files are subject to institutional data-use agreements (project LM2019-1762) and available from the corresponding author on reasonable request.
Files
Steps to reproduce
1. Download the dataset and analysis scripts from the scripts/ folder. 2. Run Python 3.13 scripts (pandas, numpy, scipy, scikit-learn) on the DEP tables to reproduce statistics and figures. 3. See README.md for full data description and study design.