Confusing FUMA Result

[Chisom Soremekun]

Hello.

I am having a little problem here. I ran FUMA using the dafault parameters and the analysis was successful.

However, the result says that i have only one lead loci but the Manhattan plot shows more peaks on different chromosomes.

Even when i created a list (76 snps) of my lead snps for FUMA to use, i still ended up getting only one lead loci.

Does anyone have any idea why this is so? Its also my first time encountering this problem?

Pictures are attached.

Thank you

[Tanya Phung]

Hello, 

For the Manhattan plot, FUMA just uses the input GWAS summary statistics that you supplied so it's independent of the downstream processes. 

For the identification of genomic loci and lead SNPs, FUMA requires that the significant SNPs (pval < 5e-8) exists in the reference. You selected 

refpanel = 1KG/Phase3
pop = AFR

First, Did you make sure that the significant SNPs in your file exist in the reference? You can check it here: https://fuma.ctglab.nl/downloadPage

Second, did you make sure that the coordinates are in hg19? 

Third, you mentioned that you supplied a lead snp file. These are the errors: 
Input lead SNP 1:41660040 does not exists in input gwas file
Input lead SNP 1:75742877 does not exists in input gwas file
Input lead SNP 1:12220138 does not exists in input gwas file
Input lead SNP 1:24960666 does not exists in input gwas file
Input lead SNP 1:15353548 does not exists in input gwas file
Input lead SNP 1:42930547 does not exists in input gwas file
Input lead SNP 1:34902776 does not exists in input gwas file
Input lead SNP 1:86133762 does not exists in input gwas file
Input lead SNP 1:24417832 does not exists in input gwas file

Input lead SNP 2:146246450 does not exists in input gwas file
Input lead SNP 2:1361345 does not exists in input gwas file
Input lead SNP 2:216700776 does not exists in input gwas file
Input lead SNP 2:144204376 does not exists in input gwas file
Input lead SNP 2:46228370 does not exists in input gwas file
Input lead SNP 2:127067636 does not exists in input gwas file

Input lead SNP 3:110016980 does not exists in input gwas file
Input lead SNP 3:65906794 does not exists in input gwas file
Input lead SNP 3:121481090 does not exists in input gwas file
Input lead SNP 3:120940977 does not exists in input gwas file

etc.... for the rest of the chromosomes

So it appears that the lead snp file might not be correct. 

Best,
Tanya

[Chisom Soremekun]

Thank you Tanya for your swift response.

I appreciate the insight on the reference panel and lead SNPs. Now I know where to start troubleshooting from.

How do I access the log file for FUMA? It will be helpful to know the outcome of the inputed lead snps.

[Chisom Soremekun]

I repeated the fuma analysis, ensuring that the list of the lead snps are in the gwas summary statistics (command below), but the result remained the same. 

Do you happen to know why this is so?

 [metat2d]$ grep -w "1:42930547" aumfuma5
1:42930547 1 42930547 t c 3.4506 6.246e-09

[Tanya Phung]

Hello, 

In my previous response, did you do this? 

I spot-checked the SNP that you mentioned above and it is not present in the reference panel. This command returned nothing:
zless 1KGphase3AFRvariants.txt.gz | awk '$1==1{print}' |  grep
 -w 42930547

A similar issue/response regarding the match to the reference panel has been answered here: https://groups.google.com/g/fuma-gwas-users/c/DSK5FAGztVI/m/PemQMyloAwAJ

Please ensure that the variants in your input GWAS summary statistics exist in the reference panel that you specify. We share the reference panel data (link above) for the 1000G reference panel. We do not have the permission to share the UKB reference panel. 

Best,
Tanya