Search results

Even though this region falls just below the significance threshold, the credible set lies entirely inside the SOX6 gene, making it quite likely that this is the relevant one. The 3D domain only includes C11orf58 as another protein coding gene. The FLAMES algorithm was also quite confident...

I'm not sure about this because these are all very small effect sizes, mutations that are common in the general, healthy population. So I see it more as small variations that aren't severe enough to cause development problems. A bit like having bigger nose or pale skin, one could also have more...

I think SHISA6 is the most likely candidate (FLAMES predicted it as well) but in case it's DNAH9, there might be a possible connection with (other weak candidates) DNAH10, and ERICH3 as they are all involved in cilia: the hair-like structures that stick out from the surface of certain cells and...

Chromatin openness Had a closer look at this region. The credible set has 25 SNPs and lies in the second intron of SHISA6 itself. There were three regulatory elements that included one of these SNPs of which the one to the right EH38E1847030 had the highest probability. It's mostly open in...

Yeah, the locus is not so clear or informative. Posting the analyses anyway, perhaps something clicks when in the future more info is available. Yes the colocalization analyses tries to find out if the two signals are driven by the same causal variant(s) based on these correlation patterns...

Here's a bit more about some of the candidate genes: CCDC92 CCDC92 Gene - GeneCards ZNF664 ZNF664 Gene - GeneCards DNAH10 DNAH10 Gene - GeneCards

Probably due to linkage disequilibrium, the correlation with the one causal SNP that is driving the signal. The colours indicate that the strips are SNPs that have the same correlation with the causal SNP.

Both the eqtl and phewas show that multiple studies that share the same signal. I've attached my results in Excel files. This region comes up a lot in metabolic traits such as BMI, HDL cholesterol, waist-to-hip ratio, etc. The ME/CFS risk version goes with a lower waist-to-hip ratio for a...

The credible set made using SUSIE consists of 77 SNPs in high correlation with each other. The region they span is full of regulatory elements which do not point to particular tissue or cell type. The 3D links data from ENCODE had multiple links to CCDC92 and ZNF664, so I suspect these are the...

This is one of the signals in DecodeME that was just below the significance threshold (top hit p = 2.43e-7). It lies in a dense region with lots of genes. The credible set lies above CCDC92 and ZNF664 and to a lesser extent DNAH10.

Another potential candidate here is ERICH3, a protein associated with cilium assembly but also neurotransmitter storage and vesicular trafficking in the brain. ERICH3: vesicular association and antidepressant treatment response - PubMed

Here's what it looks like, very similar to the ENCODE data and not supportive of NEGR1 The data file was too big we I sampled 2000 random nuclei for 7 major cell types: oligodendrocytes (ODC), VIP interneurons (Vip), D1 and D2 medium spiny neurons (MSN-D1, MSN-D2), L2/3 IT neurons (L2/3-IT)...

Haven't delved deep into what UNC13C does, it gene card says: UNC13C Gene - GeneCards

eQTL Looking at gene expression datasets, there wasn't really a result that matched the DecodeME signal well (all PP.H4 < 0.8). Phewas Looking at genetic studies of other traits, gave almost no results - very different from other loci. Two reached a PP.H4 of 0.8 or higher, dementia in the...

3D contacts with Hi-C The domains that show how the region often folds and touches in 3D points also points to the western region where UNC13C lies, and none includes the genes to the east. More info and background about this analysis is given here. The link to UNC13C in the west is also seen...

Had a closer look at this locus. The top SNP is a bit of an outlier ( lower p-value and imputation score than the SNPs it is in near perfect correlation with). So I think the signal is slightly below the significance threshold, but that doesn't matter much. chromatin openess and histone marks...

Here's what this plot looks like, it did not replicate in the Nott et al. dataset of neurons. More info about the datasets and methodology I used here. The Hi-C data from 5 ENCODE brain samples show a different picture though. No special connection to NEGR1 but to ERICH3 and LRRC53. So...

Yes. I mainly tested the neuronal cells because of the MAGMA enrichment analysis in the brain but also because POU3F2 is mainly expressed there: Tissue expression of POU3F2 - Summary - The Human Protein Atlas

Yes, I plan to do these analyses (eqtl, phewas, AlphaGenome, chromatin openness, Hi-C) on all DecodeME hits with p < 10^-6. Most of the time these analyzes don't give much extra insight, but sometimes they help refine the predictions of the FLAMES algorithm for what the most plausible causal...

Chromatin conformation I've been looking more into databases on 3-D contact: based on how the chromatin is folded, locations that are far apart on the DNA string can actually be close. It is particularly useful when the GWAS signal is in a gene desert, far away from protein-coding genes. The...