FUMA version 2.1.0 and GRCh38 support

[Tanya Phung]

Dear FUMA users, 

FUMA is now updated to version 2.1.0.

• Support for GRCh38: when you select GRCh38 option, and your input file contains chromosome and position in GRCh38 coordinates, FUMA uses dbSNP v157 to look up rsID based on the provided chromosome and position.
  • Please check https://fuma-docs.readthedocs.io/en/latest/snp2gene/quick_start.html and https://fuma-docs.readthedocs.io/en/latest/snp2gene/prepare_input_files.html for additional information. 
  • Note that this does not solve the "missing variants" problem of going from GRCh38 to GRCh37. 
• Logs: You can now download logs for your SNP2GENE jobs. Note that for this version, only the logs from the processing of the GWAS sumstat is available. Logs for other parts of SNP2GENE will be available at a later update. 

As with any updates, please expect that FUMA might be unavailable over the next few days for bug fixes. 

Thank you,
Tanya

[Frank Geller]

Dear Tanya,

Thank you for providing this useful tool.

I was delighted to see that FUMA now has GRCh38 support.

So I tried a rerun of a GWAS that previously had no issues running with hg19 positions.

I converted the old positions and extracted the six columns from the GWAS file using the recommended column header. After some trial and error with space separated files, it turned out that only a tab separated file was accepted. As you wrote somewhere else, it would be great if the GRCh38 input was as flexible as the hg19 one. But for now, it might be helpful if you add the tab separated requirement to the GRCh38 recommendations.

My hope was to have a consistent GRCh38 set of results from our GWAS and FUMA, but I can see that the positions in the results are hg19(jobID 753915).

 Can you confirm that this is the case and nothing went wrong?

Then, I’d think it would be nice to state it a bit more clearly that the GRCh38 results are only transferred to the hg19 universe (probably via the rs names?) and still analyzed based on hg19 databases.

It would be great if FUMA could provide results for GRCh38, as this is the build used by most for quite a while now.

 

All the best,

Frank

[Tanya Phung]

Hi Frank, 

Thank you for your feedback. 

The documentation (https://fuma-docs.readthedocs.io/en/latest/snp2gene/prepare_input_files.html#gwas-summary-statistics) is now updated: 

As mentioned previously (https://groups.google.com/g/fuma-gwas-users/c/E4BeXe8vox0/m/IQmJzDrQDQAJ), the issue of reference build is complicated by several factors: 

• If one wants to annotate their GWAS data entirely in GRCh38, all of the reference data that FUMA utilized would need to be in GRCh38. 
  • Some of these data (LD matrix, RNAseq, etc...) were in GRCh37 originally and would need to be either converted to GRCh38 (which is also problematic) or all resources would need to be created entirely from scratch using only those that are available in GRCh38. 
• I prefer to create an option in FUMA where users can submit files in GRCh38 and the reference databases also come from files in GRCh38 without needing to perform any conversion. This requires a lot of work and I do not have an estimate of when it will be available. 
• Given that, the currently available feature is meant to make it easier for users whose GWAS sumstats is in GRCh38 and do not want to go through the trouble of using liftover or looking up rsIDs. 

Best,
Tanya

[Frank Geller]

Hi Tanya,

thanks for the clarification. 

Of course it's not easy to find a hero doing the liftover for all the datasets.

The documentation update is very clear and should prevent further misunderstandings and some instances of ERROR 001 :)

Best,

Frank