The result shows much less number of Genomic Risk Loci than observed in the Manhattan plot

[Pronoy Kanti Mondal]

Hi Tanya,

I have a query regarding job no. 751334. The Manhattan plot generated by FUMA below appears to contain at least 8 significant loci whereas the corresponding FUMA output reports only 4 Genomic Risk Loci. 

Based on the Manhattan plot, I expected at least 8 Genomic Risk Loci in the FUMA output. Could you please clarify the reason for this discrepancy between the number of significant loci visible in the Manhattan plot produced by FUMA and the number of Genomic Risk Loci reported by FUMA?

 

Thank you

Pronoy Kanti Mondal

[Tanya Phung]

Hi Pronoy, 

The issue is that the significant SNPs that are present in the manhattan plot but not in the leadSNPs do not exist in the reference panel that you selected. 
You selected 1KG ALL: 

To check this, you can download the file below from: https://fuma.ctglab.nl/downloadPage

Then, you can verify that some of the significant SNPs exist in this reference data but others do not. For example: 

This snp exists in the database: 
zless 1KGphase3ALLvariants.txt.gz | awk '$1==7{pr
int}' | grep 144556686
7       144556686       rs567076188     G       A       0.0001997

This snp does not exist in the database: 
less 1KGphase3ALLvariants.txt.gz | awk '$1==3{print}' | grep 41976730 returns nothing meaning that this variant does not exist in the reference panel

or

less 1KGphase3ALLvariants.txt.gz | awk '$1==20{print}' | grep 62077112 returns nothing

For snps to be in the analysis for genomic risk loci, they have to exist in the reference panel. 

Best,
Tanya