GLORIA Metabolomics

Serum metabolomics analysis of fibromyalgia (FM) patients and healthy controls (HC) from the GLORIA cohort.

Overview

This repository contains the R analysis code accompanying the manuscript ​​Fibromyalgia tryptophan metabolism is associated with pain symptomology but not anti-satellite glial cell antibody levels (2026).

The analysis uses targeted metabolomics to characterize serum metabolite profiles in FM patients and healthy controls (HC).

Repository structure

.
├── GLORIA_metabolomics_analysis.Rmd   # Main analysis script
├── functions_C18.R                    # Analysis helper functions
├── func_holdouts.R                    # Cross-validation (PLS-DA) helper functions
└── README.md

Analysis steps

The GLORIA_metabolomics_analysis.Rmd script runs the full analysis in the following order:

1. Data loading — metadata, compound IDs, mass spectrometry feature data (extracted from TraceFinder)
2. Feature identification — matching features to a compound library by m/z and retention time
3. Quality control — internal standard plots, batch normalization (median normalization)
4. PLS-DA — cross-validated PLS-DA models (FM/HC, FM mild/severe, low/high IgG+SGC%) with permutation testing
5. Linear regression — age- and BMI-adjusted models for FM vs. HC
6. Spearman correlations — metabolites correlated with clinical parameters (VAS, FIQ, HAD, SF-36, PPT, CPM, PSQI, IgG+SGC%, etc.)
7. Tryptophan pathway ratios — Kyn/Trp, 5-HTrp/Trp, Ind/Trp
8. Pathway analysis — visualization of MetaboAnalyst pathway enrichment results

Requirements

R v.4.1 with the following packages:

install.packages(c(
  "tidyverse", "ggplot2", "readxl", "writexl", "stringr",
  "car", "emmeans", "pvclust", "Hmisc", "ggpubr", "plotly",
  "ggrepel", "pROC", "xml2"
))

if (!requireNamespace("BiocManager", quietly = TRUE))
  install.packages("BiocManager")
BiocManager::install(c("ComplexHeatmap", "colorRamp2", "mixOmics", "ropls"))

Data

Raw and processed data files are not included in this repository as they contain individual-level patient data. The following input files are required (paths configured via the params block in the Rmd YAML header):

Parameter	Description
metaData	GLORIA master metadata file (.xlsx)
filteredSamples	List of samples excluded due to exclusion criteria (.csv)
PSQI	PSQI sleep scores (.xlsx)
featData	Batch-normalized RPLC feature matrix (.csv)
C18ID	C18 compound reference list with functional annotation (.xlsx)
HILICID	HILIC compound library (.xlsx)
TFdt	HILIC feature data (.xlsx)
featID	Additional IDs (.csv)
funcAnn	Manual annotation table (.xlsx)

Output

Results are written to the directory specified by params$outputDir, including:

• tables/ — metadata summary table, full results, supplementary tables
• FC/ — fold-change heatmaps and volcano plots
• correlation/ — metabolite–clinical parameter correlation heatmaps
• PLSDA/ — PLS-DA scores, loadings, VIP plots, and performance metrics
• publication/ — final figures and tables for the manuscript

Contact

Jenny Jakobsson