ERROR: 5 (Error from lead SNPs and candidate SNPs identification / No significant SNPs were identified)

[Wisdom A Akurugu]

Hi FUMA Managers

I submitted a file for analysis. I have read all the comments on this forum related to it and made all the corrections but it has not resolved my problem.

Job ID:258715

My rsIDs and positions match that of dbSNP.

I will be grateful for your assistance

Regards

Wisdom

[d.p.wi...@vu.nl]

Hi Widsom, 

Based on the logs, it looks like there were only two variants correctly processed by FUMA and neither of those variants were input lead SNPs. I am guessing your file is not formatted correctly causing lots of variants to be lost. Below is the information for the log file. 

1 SNPs in chromosome 4
Input lead SNP rs113631626 does not exists in input gwas file
Start chromosome 11 ...
1 SNPs in chromosome 11
Input lead SNP rs373351583 does not exists in input gwas file

Can you share the first few lines of your input file?

Cheers,

Doug

[Wisdom A Akurugu]

Hi Doug,

Thanks for the reply.

I had intentionally made the input file to contain only 2 variants and the leads file to contain the same 2 varints as below:

Input:

CHR BP P A1 A2 BETA SE

4 106156140 5.49E-08 A G 0.1146 0.01894
11 117782607 1.50E-07 T C 0.0871 0.01501

Leads:

rsID CHR BP P A1 A2 BETA SE

rs113631626 4 106156140 5.49E-08 A G 0.1146 0.01894
rs373351583 11 117782607 1.50E-07 T C 0.0871 0.01501

There are many of them that do not get annotated and I had to pick just these 2 variants to try and see what the problem is.

I will be happy for your timely reply

Regards

Wisdom

[d.p.wi...@vu.nl]

Hi Wisdom, 

rs113631626 is not in the reference panel, likely because it is very rare https://www.ncbi.nlm.nih.gov/snp/rs113631626 

rs373351583 in our data does not match the position you gave (11:117782606:C:T rs369575036) but your position does match the dbSNP data https://www.ncbi.nlm.nih.gov/snp/?term=rs373351583. I am not sure why our reference data has the wrong position for this SNP. I will look into this but it may be due to an update in the new version of dbSNP that is not reflected in our data.

Cheers,

Doug

[Wisdom A Akurugu]

Hi Doug,

Thanks for the quick response. I have really be stuck with this issue for some weeks.

I assume that most of my variants that were dropped is probably the rarity/missing variants in your panel relative to mine or the position shift.

Will it be fine with you that I compile them from by side and send to you as an email to crosscheck for the compatibility? That could be affecting a number of other users.

Please let me know.

Regards

Wisdom

[d.p.wi...@vu.nl]

Hi Wisdom, 

You can download the reference panel SNP list here https://fuma.ctglab.nl/tutorial#refpanel and compare your input data to our reference panel.

Cheers,

Doug

[Wisdom A Akurugu]

Thank Doug. I will compare my variants and let you know

[Wisdom A Akurugu]

Hi Doug,

I have gone through your reference and found that were a number of variants in my data with the correct dbSNP IDs and chrom positions but not in your reference. Are they considered rare variants from your side?

rsID                     CHR              BP                   dbSNP
rs367763945     11             48285879       11:48285879 (GRCh37)
rs769016513     11             48286074       11:48286074 (GRCh37)
rs1257441095   11             48510981       11:48510981 (GRCh37)
rs369084263     13             103268814     13:103268814 (GRCh37)
rs565127215     5                102509631     5:102509631 (GRCh37)
rs113631626     4               106156140      4:106156140 (GRCh37)
rs149086811***     8                3889521          8:3889521 (GRCh37)
rs373351583    11              117782607      11:117782607 (GRCh37)

Is possible to include them in your reference?

The variant asterisked does not have an rsID in your reference but it is a variant.

Kind regards

Wisdom

[d.p.wi...@vu.nl]

Hi Wisdom, 

Unfortunately not, the data we have is from the 1000 genomes project and we cannot add variants that are not present in this data. We need these variants to be present in the people included in the 1000 genomes data so we can get linkage disequilibrium estimates for them. Unfortunately, the variants not present in the reference panel will be ignored when creating genomic risk loci in FUMA.

I am not sure why these variants are not present in the people included in the 1000 genomes data, them being rare is one possibility. 

Cheers,

Doug

[Rui Marçalo]

Hi Doug,

I am also getting error 5 (jobID: 276931) on FUMA.

I have a significant SNP in the imput file (p=4.778e-8) however it is not being recognized by FUMA.

I have checked the that the variant is present in the 1000G (EUR) project and that it has the same position (5167688) and effect allele (GTT) as in my input file.

I have set the MAF threshold parameter to 0, so that should not be the reason.

I have also tried providing the rsID (both the current and the one in the 1000G project (which has been deprecated)) but the error persisted.

Could you please give me some guidance on how to solve this? 

7 118504249:A:C 118504249 C ADD 598 3.271 4.536 6.934e-06
7 118510105:GA:G 118510105 G ADD 598 3.271 4.536 6.934e-06
7 118512118:A:C 118512118 C ADD 598 3.209 4.486 8.726e-06
8 5165998:C:T 5165998 C ADD 598 3.195 4.661 3.894e-06
8 5166336:T:C 5166336 T ADD 598 3.195 4.661 3.894e-06
8 5166353:T:C 5166353 T ADD 598 3.195 4.661 3.894e-06
8 5166615:A:G 5166615 A ADD 598 3.215 4.69 3.391e-06
8 5166623:T:C 5166623 T ADD 598 3.144 4.585 5.551e-06
8 5167168:A:T 5167168 A ADD 598 3.195 4.661 3.894e-06
8 5167333:G:A 5167333 G ADD 598 3.195 4.661 3.894e-06
8 5167688:GT:G 5167688 GT ADD 598 3.195 4.661 3.894e-06
8 5167688:GT:GTT 5167688 GTT ADD 598 4.174 5.531 4.778e-08
8 5168044:T:A 5168044 T ADD 598 3.195 4.661 3.894e-06
8 5168886:G:A 5168886 G ADD 598 3.101 4.517 7.563e-06

Thank you

Rui

[d.p.wi...@vu.nl]

Hi Rui,

Can you try adding a column to your input data that specifies the non-effect allele? Here it would be GT.

Cheers,

Doug

[Rui Marçalo]

Hi Doug,

I am having trouble finding a way to test your suggestion.

I can't figure out how to run the GWAS using PLINK and have the non-effect allele in the output file.

Do you have any suggestion on how to do this?

Maybe it is very simple and I am just not finding it.

Thanks

Rui

[d.p.wi...@vu.nl]

Hi Rui,

You can merge your summary statistics file with the plink .bim file and select the allele that is not the tested allele from the .bim file. I would do this using R but there are many ways to do this.

Cheers,

Doug