SNP2GENE questions

[Jamee Berg]

Hi! I ran a SNP2GENE run on 2026_06_05. I have two questions

1. Is the annov.stats results file different from what is displayed in the “Functional consequences of SNPs on genes” graph in the “summary of results?” The numbers do not match from the two.

 

2. my xQTL output file “xqtls.txt” has 1070786 rows and the columns are:

colnames(xqtls)

[1] "uniqID"            "db"                "tissue"          

[4] "protein"           "type"              "qtl_type"         FU

[7] "genomicriskloci"   "ensemble_id"       "originalPhenotype"

However, there is no p-value. Can you help me interpret these results?

Thank you, Jamee

[Tanya Phung]

Hi Jamee, 

You did not provide a jobID, but I think you are referring to your SNP2GENE job 742040, is that correct? 

Could you elaborate more on which numbers do not match? I spot-checked a few and it seemed to match. For example: 

Intronic: 

grep -we "intronic" -e annot annov.stats.txt | awk '{print$1"\t"$4"\t"$5"\t"$6"\t"$7}'
annot   count   prop    enrichment      fisher.P
intronic        5133    0.425269262634631       1.16729752720614        3.72503112240164e-43

Upstream:

grep -we "upstream" -e annot annov.stats.txt | awk '{print$1"\t"$4"\t"$5"\t"$6"\t"$7}'
annot   count   prop    enrichment      fisher.P
upstream        133     0.0110190555095278      1.03633686951234        0.65713856139728

Splicing:

grep -we "splicing" -e annot annov.stats.txt | awk '{print$1"\t"$4"\t"$5"\t"$6"\t"$7}'
annot   count   prop    enrichment      fisher.P
splicing        3       0.000248550124275062    2.20284705531045        0.157320617524645

For your second question, in this xQTL mapping analysis, the datasets were already pre-processed to contain only the significant variant-gene pair. So for example, in eQTLs from GTEx, the datasets that FUMA uses only contain the significant association between the variant and the genes. How the significance is determined is from the original studies This is true for all but 2 datasets. You can read more here: https://github.com/tanyaphung/fuma_qtls. Therefore, there is no p values here because this is simply just asking if a GWAS hit was annotated as an QTL previously, using the threshold defined by the original's study. 
So in this analysis, what FUMA does is that for each of the GWAS hits, it checks if this GWAS hit is an QTL in any of the selected datasets. In this same job 742040 you selected 108 datasets, so there may be a high number of GWAS hits that are QTLs. It could be that the same genes are coming up in different tissues and cell types. 
For more information, perhaps you can check this: https://fuma-docs.readthedocs.io/en/latest/practicals/scz_gwas_2022.html#run-a-snp2gene-job-with-xqtls-mapping

I hope this helps,
Tanya