Skip to content

Chr1 173.5M-174.5M

Methodology

To narrow the DecodeME1 GWAS-1 signal, I fine-mapped the hit on chromosome 1 using SUSIE2.

As a linkage disequilibrium reference, I used a UK Biobank LD matrix hosted on AWS Open Data. Because this LD reference uses GRCh37 coordinates, I used GWASLab to liftover the DecodeME summary statistics to GRCh37.

As a sensitivity analysis, I ran SUSIE 4 times:

  • Once with \(L=10\),
  • Once with \(L=2\),
  • Once with \(L=1\),
  • Once with \(L=10\) and strict variant filtering.

\(L\) refers to the maximum number of credible sets that can found by SUSIE. A lower \(L\) corresponds to increased regularization, since it decreases the ability of SUSIE to use extra credible sets to fit noise. Weissbrod et al.3 observe that setting \(L\) to 1 protects against mismatch between the LD reference population and the GWAS population, because when \(L=1\), SUSIE no longer depends on the LD matrix. They also observe that when \(L=2\), even though SUSIE still depends on the LD matrix, empirically it tends to be robust to moderate levels of population mismatch. I thus used the \(L=1\) and \(L=2\) runs to evaluate whether population mismatch could be influencing SUSIE's results.

"Variant filtering" refers to removal of outlier variants according to a Kriging-based likelihood ratio test. Zou et al.4 propose this filtering strategy to mitigate instability in SUSIE due to mismatch between the LD and GWAS populations. In the first three runs above, I filter variants with a likelihood ratio (\(\mathrm{LR}\)) and absolute \(z\) score greater than 2, consistent with the SUSIE documentation. In the final run I instead filter variants with \(\mathrm{LR}\ge 2\) and \(|z|\ge 1\), to evaluate the sensitivity of the results to the filtering threshold.

In my SUSIE runs, I retained palindromic SNPs whose strand orientation GWASLAB was able to determine from allele frequencies in the Thousand Genomes Project, and discarded other palindromic SNPs.

Results

In all 4 runs, SUSIE found a single diffuse credible set. Moreover, this credible set contained the same 86 variants in all four runs, as illustrated in the UpSet plot below:

upset_chrom_1

The next figure illustrates the SUSIE results for \(L=10\). It is representative of the other runs.

chr1_stackplot

  • The top panel is a heatmap in which pixel \((i,j)\) is colored according to the squared correlation between variants \(i\) and \(j\). The heatmap reveals the local linkage disequilibrium (LD) structure in the vicinity of the GWAS hit, which is a determinant of SUSIE's results when \(L>1\).

  • The second panel shows a local Manhattan plot.

  • The third panel shows the SUSIE posterior inclusion probability (PIP).

  • The bottom panel shows genes in the region of the GWAS hit.

Overall, SUSIE has returned a diffuse signal in a region with a number of plausible genes. This makes it unclear which genes deserve follow-up investigation.

The table below lists the full detailed SUSIE results for the \(L=10\) case

Variant List
cs CHR POS EA NEA alpha mu PIP
L1 1 173815290 C T 0.0349532 -0.022181 0.0349532
L1 1 173853127 C T 0.0334384 -0.0221481 0.0334384
L1 1 173865586 T C 0.0332454 -0.0221439 0.0332454
L1 1 173815111 C T 0.0332354 -0.0221436 0.0332354
L1 1 173866074 A T 0.0332311 -0.0221435 0.0332311
L1 1 173878862 C T 0.0328582 -0.0221352 0.0328582
L1 1 173812639 A C 0.0327682 -0.0221331 0.0327682
L1 1 173851310 A G 0.0313245 -0.0220996 0.0313245
L1 1 173859100 G A 0.0296411 -0.0220585 0.0296411
L1 1 173863209 A G 0.0296014 -0.0220575 0.0296014
L1 1 173863569 A T 0.0292311 -0.0220482 0.0292311
L1 1 173863567 T G 0.0292311 -0.0220482 0.0292311
L1 1 173846590 G T 0.0276321 -0.0220062 0.0276321
L1 1 173863568 T A 0.0275657 -0.0220044 0.0275657
L1 1 173832336 T C 0.0272003 -0.0219944 0.0272003
L1 1 173878832 C T 0.0261331 -0.0219645 0.0261331
L1 1 173838788 T TG 0.0261228 -0.0219642 0.0261228
L1 1 173855298 T A 0.0255997 -0.0219491 0.0255997
L1 1 173846110 A G 0.025353 -0.0219418 0.025353
L1 1 173857283 G A 0.0243066 -0.0219102 0.0243066
L1 1 173848009 G A 0.023978 -0.0219 0.023978
L1 1 173824813 T C 0.0234489 -0.0218833 0.0234489
L1 1 173842467 G A 0.0230966 -0.0218719 0.0230966
L1 1 173870321 G GTAC 0.0230538 -0.0218705 0.0230538
L1 1 173857037 T C 0.0228147 -0.0218627 0.0228147
L1 1 173881871 T C 0.0203149 -0.0217753 0.0203149
L1 1 173844051 T A 0.0184477 -0.0217024 0.0184477
L1 1 173820365 C T 0.0145082 -0.0215198 0.0145082
L1 1 173832772 CA C 0.0144368 -0.021516 0.0144368
L1 1 173878471 G A 0.0118415 -0.0213641 0.0118415
L1 1 173831882 G A 0.00912575 -0.0211628 0.00912575
L1 1 173743879 CAAAA C 0.00696391 -0.0209519 0.00696391
L1 1 173717200 ACT A 0.00675049 -0.0209275 0.00675049
L1 1 173767443 T A 0.00660937 -0.0209109 0.00660937
L1 1 173783493 C T 0.00650478 -0.0208983 0.00650478
L1 1 173699007 G A 0.00592378 -0.0208246 0.00592378
L1 1 173698510 T C 0.00585287 -0.0208151 0.00585287
L1 1 173709616 G T 0.0056909 -0.020793 0.0056909
L1 1 173734270 CAACA C 0.0056717 -0.0207903 0.0056717
L1 1 173683954 T C 0.00516697 -0.0207165 0.00516697
L1 1 173755936 TGAAG T 0.00476139 -0.0206516 0.00476139
L1 1 174210076 T TTG 0.00289739 -0.0202525 0.00289739
L1 1 174128994 G A 0.00239407 -0.0200971 0.00239407
L1 1 174158856 C T 0.00235282 -0.0200829 0.00235282
L1 1 174111115 T C 0.00229859 -0.0200638 0.00229859
L1 1 174066947 T C 0.00227798 -0.0200564 0.00227798
L1 1 174069469 CT C 0.00227531 -0.0200555 0.00227531
L1 1 174084104 A G 0.0022655 -0.0200519 0.0022655
L1 1 174152688 G A 0.00225032 -0.0200464 0.00225032
L1 1 174146656 C A 0.00222885 -0.0200385 0.00222885
L1 1 174085043 TACA T 0.00219654 -0.0200266 0.00219654
L1 1 174076864 C G 0.00218651 -0.0200228 0.00218651
L1 1 174064481 A T 0.0021761 -0.0200189 0.0021761
L1 1 174068049 A G 0.00215682 -0.0200116 0.00215682
L1 1 174069981 C T 0.00215307 -0.0200102 0.00215307
L1 1 174191694 C T 0.00214725 -0.0200079 0.00214725
L1 1 174062911 C T 0.00214307 -0.0200063 0.00214307
L1 1 174057626 T C 0.00214136 -0.0200057 0.00214136
L1 1 174241165 C G 0.0021385 -0.0200046 0.0021385
L1 1 174093837 G T 0.00213651 -0.0200038 0.00213651
L1 1 174157834 A G 0.00213405 -0.0200029 0.00213405
L1 1 174170341 C T 0.00213381 -0.0200028 0.00213381
L1 1 174086292 T C 0.00211439 -0.0199953 0.00211439
L1 1 174050667 G A 0.00210878 -0.0199931 0.00210878
L1 1 174068929 A G 0.00207848 -0.0199812 0.00207848
L1 1 174079700 G C 0.00207189 -0.0199786 0.00207189
L1 1 174118271 C G 0.00206586 -0.0199762 0.00206586
L1 1 174075924 G T 0.00206197 -0.0199746 0.00206197
L1 1 174075925 C T 0.00206197 -0.0199746 0.00206197
L1 1 174059522 C G 0.002049 -0.0199694 0.002049
L1 1 174161430 A G 0.00203663 -0.0199645 0.00203663
L1 1 174072203 G A 0.00203588 -0.0199642 0.00203588
L1 1 174078417 A G 0.00202943 -0.0199615 0.00202943
L1 1 174079584 T A 0.00202752 -0.0199608 0.00202752
L1 1 174079585 G A 0.00202752 -0.0199608 0.00202752
L1 1 174069421 T C 0.00191843 -0.0199152 0.00191843
L1 1 174176363 G A 0.00190403 -0.019909 0.00190403
L1 1 174144293 CACAT C 0.00189623 -0.0199056 0.00189623
L1 1 174197408 G A 0.00187952 -0.0198983 0.00187952
L1 1 174078860 C T 0.00187677 -0.0198971 0.00187677
L1 1 174113408 G A 0.0018707 -0.0198944 0.0018707
L1 1 174152476 C T 0.0018538 -0.0198869 0.0018538
L1 1 174150850 C T 0.00182508 -0.019874 0.00182508
L1 1 174060062 G A 0.00180084 -0.0198629 0.00180084
L1 1 174326702 G A 0.00161599 -0.0197731 0.00161599
L1 1 174288602 G A 0.00158912 -0.0197592 0.00158912

  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. Gao Wang, Abhishek Sarkar, Peter Carbonetto, and Matthew Stephens. A simple new approach to variable selection in regression, with application to genetic fine mapping. Journal of the Royal Statistical Society Series B: Statistical Methodology, 82(5):1273–1300, 2020. URL: https://academic.oup.com/jrsssb/article/82/5/1273/7056114

  3. Omer Weissbrod, Farhad Hormozdiari, Christian Benner, Ran Cui, Jacob Ulirsch, Steven Gazal, Armin P Schoech, Bryce Van De Geijn, Yakir Reshef, Carla Márquez-Luna, and others. Functionally informed fine-mapping and polygenic localization of complex trait heritability. Nature Genetics, 52(12):1355–1363, 2020. URL: https://pmc.ncbi.nlm.nih.gov/articles/PMC7710571/

  4. Yuxin Zou, Peter Carbonetto, Gao Wang, and Matthew Stephens. Fine-mapping from summary data with the “Sum of Single Effects” model. PLoS Genetics, 18(7):e1010299, 2022. URL: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010299