Significant genes from gene-property analysis in SNP2GENE

[Surati Kumari]

Hello 

I used the BrainSpan dataset for gene-property analysis in SNP2GENE module, which returns "magma_exp_*_log2RPKM.gsa.out". Although this file gives overall idea of which grouped labels are found to have significant association with the trait in the Brainspan dataset, it fails to provide a list of genes with significant overlapping association with the trait and the gene expression in a specific label/category. For example, I got 8_pcw in Brainspan dataset with significant association, however I also want to know which genes from that age category are showing the positive relationship. 

Is it possible to derive the required information somehow?  I have downloaded the Brainspan dataset 'genes_matrix_csv.zip', however its not pre-processed as per FUMA yet. Would I have to run magma locally with the command mentioned in "Gene property analysis for tissue specificity" section in FUMA SNP2GENE tutorial.

Can I just pass magma.genes.out/.raw generated from SNP2GENE as input file in GENE2FUNC? Will it possibly lead to the required result?

Thanks in advance

Surati

[Tanya Phung]

Hi Surati, 

I am not 100% certain what your question is, but are you asking: "what genes drive the significant MAGMA tissue expression in this list?" If so, running MAGMA gene property analysis as implemented in FUMA might not be suitable for your question because in MAGMA gene property, the whole idea of the gene property analysis is that we do not want to determine an a priori threshold because there might be sub-threshold information. 

You may want to check out a paper from my colleague recently (https://www.medrxiv.org/content/10.1101/2025.10.10.25337470v1) where he "prioritised genes that had significant GWAS P and were in the top 10% of overexpression, where overexpression was calculated based on log2fc of tissue/cell expression vs average." 

Hope that helps,
Tanya