Skip to content

MAGMA DecodeME Analysis

MAGMA GTEx Analysis

As an initial step, I applied MAGMA to DecodeME, partially reproducing analysis from the DecodeME preprint1.

MAGMA Gene Analysis

I applied MAGMA's SNP-wise-mean model to the summary statistics from DecodeME's GWAS-1.

In this step:

  • Data from the 1000 genomes projects was downloaded from the MAGMA website and used as a linkage disequilibrium reference.
  • Build 151 of dbSNP was used to assign RSIDs to SNPs.
  • Magma's default proximity-based rules were used to assign SNPs to genes.

MAGMA produces a table of genes, effect sizes, and p values. Filtering these genes via the Benjamini-Hochberg procedure2 at a false discovery rate of 0.01, and joining with a database of gene descriptions from Ensembl Biomart produces the following table:

GENE Gene name CHR P Gene description
ENSG00000033122 LRRC7 1 1.724e-09 leucine rich repeat containing 7 [Source:HGNC Symbol;Acc:HGNC:18531]
ENSG00000124214 STAU1 20 7.0767e-09 staufen double-stranded RNA binding protein 1 [Source:HGNC Symbol;Acc:HGNC:11370]
ENSG00000124207 CSE1L 20 3.2463e-08 chromosome segregation 1 like [Source:HGNC Symbol;Acc:HGNC:2431]
ENSG00000135090 TAOK3 12 2.3082e-07 TAO kinase 3 [Source:HGNC Symbol;Acc:HGNC:18133]
ENSG00000124198 ARFGEF2 20 3.1435e-07 ARF guanine nucleotide exchange factor 2 [Source:HGNC Symbol;Acc:HGNC:15853]
ENSG00000117593 DARS2 1 3.3379e-07 aspartyl-tRNA synthetase 2, mitochondrial [Source:HGNC Symbol;Acc:HGNC:25538]
ENSG00000185278 ZBTB37 1 4.503e-07 zinc finger and BTB domain containing 37 [Source:HGNC Symbol;Acc:HGNC:28365]
ENSG00000250091 DNAH10OS 12 6.5098e-07 dynein axonemal heavy chain 10 opposite strand [Source:HGNC Symbol;Acc:HGNC:37121]
ENSG00000179195 ZNF664 12 1.1217e-06 zinc finger protein 664 [Source:HGNC Symbol;Acc:HGNC:25406]
ENSG00000119242 CCDC92 12 1.2073e-06 coiled-coil domain containing 92 [Source:HGNC Symbol;Acc:HGNC:29563]
ENSG00000197935 ZNF311 6 1.2929e-06 zinc finger protein 311 [Source:HGNC Symbol;Acc:HGNC:13847]
ENSG00000158406 H4C8 6 1.8071e-06 H4 clustered histone 8 [Source:HGNC Symbol;Acc:HGNC:4788]
ENSG00000198216 CACNA1E 1 2.2062e-06 calcium voltage-gated channel subunit alpha1 E [Source:HGNC Symbol;Acc:HGNC:1392]
ENSG00000187323 DCC 18 2.2395e-06 DCC netrin 1 receptor [Source:HGNC Symbol;Acc:HGNC:2701]
ENSG00000111707 SUDS3 12 2.4987e-06 SDS3 homolog, SIN3A corepressor complex component [Source:HGNC Symbol;Acc:HGNC:29545]
ENSG00000197653 DNAH10 12 3.4281e-06 dynein axonemal heavy chain 10 [Source:HGNC Symbol;Acc:HGNC:2941]
ENSG00000089220 PEBP1 12 3.9196e-06 phosphatidylethanolamine binding protein 1 [Source:HGNC Symbol;Acc:HGNC:8630]
ENSG00000188730 VWC2 7 5.7778e-06 von Willebrand factor C domain containing 2 [Source:HGNC Symbol;Acc:HGNC:30200]
ENSG00000028116 VRK2 2 7.4569e-06 VRK serine/threonine kinase 2 [Source:HGNC Symbol;Acc:HGNC:12719]
ENSG00000117601 SERPINC1 1 8.7656e-06 serpin family C member 1 [Source:HGNC Symbol;Acc:HGNC:775]

MAGMA Gene Property Analysis

I next applied MAGMA's gene property analysis module to DecodeME. This step combined the gene analysis results above with tissue-specific RNA expression values from the GTEx project3. The aim was to identify tissues enriched for genes associated with ME/CFS. The results are shown in the bar plot below:

bar_plot_decodeme_tissues In this plot, the y-axis corresponds to \(-\log_{10}(p)\) values, the x-axis corresponds to tissue type (only tissues with the smallest p values are shown), and bars are colored according to whether their p value meets Bonferroni-corrected significance threshold,

These results unambiguously point to the nervous system as a major site of ME/CFS gene activity.

Reproducing

To reproduce this analysis, run the DecodeME Initial Analysis Script.

Follow-Up Questions

  1. Do other approaches to identify significant tissues from GWAS-summary statistics produce concordant or discordant results?
  2. How reliable is the GTEx-based MAGMA gene-set-analysis approach to identifying significant tissues? In other words: for diseases with well-understood pathological processes, does it produce results consistent with these processes?

  1. Genetics Delivery Team, Thibaud Boutin, Andrew D Bretherick, Joshua J Dibble, Esther Ewaoluwagbemiga, Emma Northwood, Gemma L Samms, Veronique Vitart, Project, Cohort Delivery Team, Øyvind Almelid, and others. Initial findings from the DecodeME genome-wide association study of myalgic encephalomyelitis/chronic fatigue syndrome. medRxiv, pages 2025–08, 2025. URL: https://www.medrxiv.org/content/10.1101/2025.08.06.25333109v1. ↩

  2. Yoav Benjamini and Yosef Hochberg. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological), 57(1):289–300, 1995. URL: https://rss.onlinelibrary.wiley.com/doi/abs/10.1111/j.2517-6161.1995.tb02031.x. ↩

  3. GTEx Consortium. The GTEx consortium atlas of genetic regulatory effects across human tissues. Science, 369(6509):1318–1330, 2020. URL: https://www.science.org/doi/full/10.1126/science.aaz1776. ↩