1) ERROR 005 and 2) table loading error / website logs me out.

[luba....@c4xdiscovery.com]

Hello

I was hoping you could help with my query. When testing FUMA's SNP2Gene tool with various SNPs, I noticed a couple of issues:

1) ERROR 005: When testing various SNPs (using a .txt file, as per the online instructions), I noticed that for one SNP, rs143918452 (Job IDs: 16843, 16844, 16846, 16847), SNP2Gene tool returns an error (ERROR 005). The text file I used contained a single SNP: rs143918452.

When I run a different file, containing a number of different SNPs, including rs143918452 (Job ID 16850), I get results for all SNPs, except rs143918452. In the results, there is no message saying that no analysis was performed for rs143918452.

2) Sometimes the website gives off table loading errors (see my screen shots attached). The website also occasionally logs me out soon after I log in, showing exactly the same table loading errors.

I would be very grateful for and look forward to your response.

Best wishes

Luba Prout

[Kyoko Watanabe]

Hi Luba,

Thank you for using FUMA and I am sorry bout the inconvenience.

For point 1, rs143918452 has a MAF < 0.01 in 1KG Phase3 EUR panel. Since the default MAF threshold for FUMA is 0.01, and FUMA uses MAF of reference panel for filtering, this SNP did not survive after filtering.
Please set MAF threshold to 0 when you submit job so that you don't miss this SNP.

For point 2, I am aware of that there are some issues with authorization. Often if you clear the local cache and log in again you should be fine. But I will try to improve this.

Best,
Kyoko

[luba....@c4xdiscovery.com]

Thank you very much for the quick response and clarifications!

Luba

[luba....@c4xdiscovery.com]

Hi Kyoko

I am hoping you could help with another quick query. I have run a number of jobs to test more SNPs and there is another issue.

I changed MAF to 0, and ran different files with mixtures of SNPs (some SNPs are repeating in the different jobs), job IDs: j19-Opt1 to j19-Opt8. I noticed that output for some SNPs is missing in some jobs but is present in others. Results for some SNPs are missing in all the runs, with MAF = 0. All SNPs are GRCh37 and GRCh38 compatible.

Looking forward to hearing from you.

Best wishes

Luba

P.S. Great tool, thanks for making it available!

[Kyoko Watanabe]

Hi Luba,

Could you please specify which SNPs you are missing in which jobID?
And it'd be very appreciated if you can give me your jobID not the job title.
Other than MAF, there is also filtering based on P-value (by default 0.05).
Are you sure this is not the case for you?

Best,
Kyoko

[luba....@c4xdiscovery.com]

Hi Kyoko

Thanks for the quick response and apologies - here are the job IDs: 17045, 17046, 17059, 17063, 17073, 17087.

The parameters were identical in all these jobs; the only difference was the input file. In each input file, there was a list of SNPs, all with P value = 1E-18.

I ran a number of different SNPs but just to highlight, here is an example:

In the job ID 17045, the output for rs3829628 was missing.

In the job ID 17046, the output for rs7014166 and rs6581338 was missing.

In the job ID 17059, again with identical parameters, I ran the above-mentioned three SNPs: rs3829628, rs7014166 and rs6581338 (all together), and with other SNPs, and the job returned output for rs3829628, but not the other two SNPs.

In the subsequent two jobs (IDs: 17063, 17073), I again ran rs7014166 and rs6581338 (with other SNPs). The output for rs6581338 was present in job ID 17073, but not in 17063; there were no results for rs7014166 in either job ID 17063 or 17073. I ran rs7014166 again (job ID 17087) but no output was present for this SNP.

All the above-mentioned SNPs were present in all the relevant input files.

I look forward to hearing from you.

Luba

[Kyoko Watanabe]

Hi Luba,

I apologies for the delay, I was out of office last two days.

I'm not sure if I understand correctly what you meant by "missing SNPs".
I checked the SNPs with the jobID and I can see all of the SNPS you mentioned exist in the jobs you specified.
For example, there is rs3829628 in job 17045, and all rs3829628, rs7014166 and rs6581338 exist in job 17059.
Did you check the SNP table (snps.txt)?
Or do you mean "missing" by those SNPs do not exist in risk loci file?
It's very well possible that those SNPs are not top SNPs but in LD with other top SNPs.
In that case you won't find those SNPs in the risk loci table but in SNPs table

Is this clear?

Best,
Kyoko

[luba....@c4xdiscovery.com]

Hi Kyoko

Again, thanks for your reply.

In my previous email, I should have specified that the output I was referring to was in relation to mapped genes. In some instances the SNPs mentioned were mapped to genes and in others weren't, despite the fact that all parameters were identical. I am sorry I am new to both the field and FUMA and so would really appreciate your feedback on why this might be the case.

Thank you.

Luba

[Kyoko Watanabe]

Hi Luba,

Ah, ok, I'm sorry it might be bit confusing but IndSigSNPs column in the gene table is only the independent significant SNPs but not the SNPs mapped to the gene, since if I list actually mapped SNPs in that table, that column will be huge.
So for example, if a gene A has SNP rsXXX in the IndSigSNPs column, that means there are SNPs mapped to gene A which are in LD with rsXXX but rsXXX does not necessary mapped to the gene.

Hope it's clear and no worries, I'm always happy to help understanding FUMA results;)

Best,
Kyoko

[luba....@c4xdiscovery.com]

Hello again

Thank you for the clarification.

Ok, but why is there a difference between the jobs (where the same parameters are used):


job ID 17045

rs3829628
genomic risk loci - absent
lead SNPs - absent
Ind Sig SNPs - absent
mapped genes - absent


job ID 17046

rs6581338
genomic risk loci - absent
lead SNPs - absent
Ind Sig SNPs - absent
mapped genes - absent


job ID 17059

rs3829628
genomic risk loci - present
lead SNPs - present
Ind Sig SNPs - present
mapped genes - present

rs6581338
genomic risk loci - absent
lead SNPs - absent
Ind Sig SNPs - absent
mapped genes - absent


job ID 17063

rs6581338
genomic risk loci - absent
lead SNPs - absent
Ind Sig SNPs - absent
mapped genes - absent


job ID 17073

rs6581338
genomic risk loci - present
lead SNPs - present
Ind Sig SNPs - present
mapped genes - present

Is there some kind of background filtering?

Best wishes

Luba

[Kyoko Watanabe]

Hi Luba,

That's the difference of input summary statistics.
I don't know what you exactly used for input (since I don't have access to user inputs).
But here is an example (and I assume you understand how clumping works).
If you don't include any SNPs with P-value less that the P-value of rs3829628 and in LD with rs3829628.
In that case, rs3829628 will be the top SNP in its LD block since there is no other SNPs with less P-value.
When you include other SNPs with P-value less that rs3829628, top SNP will change (as top SNP has the minimu P-value in a LD block).
If you don't understand this, maybe you should look up what the clump is?

Best,
Kyoko

[luba....@c4xdiscovery.com]

Hi Kyoko

Thanks for the quick response!

In the input files, the p value for all SNPs = 1E-18.

Luba