hanchenphd/GMMAT

Error in eval(parse(text = x, keep.source = FALSE)[[1L]]) : object 'SNP__' not found

Closed this issue · 5 comments

dylkot commented

Hi There,

Thanks again for this great tool and apologies for a second question. I am now trying to run Wald test on a subset of the imputed dosages in order to get ORs. My command is:

res <- glmm.wald(formula, data = pheno, kins = rel, id = "IID", random.slope = NULL, groups = NULL,
                 family = binomial(link = "logit"), infile = infile,  snps=snps, is.dosage = TRUE, verbose = TRUE,
                 missing.method="impute2mean", snp.col=1, infile.ncol.print=1, 
                 infile.ncol.skip=1, infile.sep='\t', infile.nrow.skip=1, infile.header.print="SNP")

and the first 5 rows/columns of the input data input data looks like

ID	SID1	SID2	SID3	SID4
1:1917446[b37]G,A	0	0	0	0.85
1:1919318[b37]C,A	0	0	0	0.85
1:3498779[b37]C,T	0	0	0	0
1:3509544[b37]G,T	0	0	0	0
1:20479748[b37]C,T	0	0	0	0

snps is a vector like

c("1:1917446[b37]G,A"  "1:1919318[b37]C,A"  "1:3498779[b37]C,T"  "1:3509544[b37]G,T"  "1:20479748[b37]C,T")

and I am getting the following error message:

Progress of Wald test:
  |                                                                                                                                                                                               |   0%

Analyze SNP  1 :  1:1917446[b37]G,A 
Error in eval(parse(text = x, keep.source = FALSE)[[1L]]) : 
  object 'SNP__' not found
In addition: Warning message:
In glmm.wald(formula, data = pheno, kins = rel, id = "IID", random.slope = NULL,  :
  Argument select is unspecified... Assuming the order of individuals in infile matches unique id in data...

Any idea what the cause of the bug is?

dylkot commented

As an alternative approach, I load the data in from a vcf.gz file like so

gdsfn <- paste(invcf, ".gds", sep="")
SeqArray::seqVCF2GDS(invcf, gdsfn, parallel=TRUE, fmt.import="DS", scenario=c("imputation"))
snps = seqGetData(gdsfn, "variant.id")
res <- glmm.wald(formula, data = pheno, kins = rel, id = "IID", random.slope = NULL, groups = NULL,
                 family = binomial(link = "logit"), infile = gdsfn,  snps=snps, is.dosage = TRUE, verbose = TRUE,
                 missing.method="impute2mean")

and in this case, the test runs but just returns all NAs for all output values.

 SNP CHR POS REF ALT  N AF BETA SE PVAL converged
1   1  NA  NA  NA  NA NA NA   NA NA   NA        NA
2   2  NA  NA  NA  NA NA NA   NA NA   NA        NA
3   3  NA  NA  NA  NA NA NA   NA NA   NA        NA
4   4  NA  NA  NA  NA NA NA   NA NA   NA        NA
5   5  NA  NA  NA  NA NA NA   NA NA   NA        NA
6   6  NA  NA  NA  NA NA NA   NA NA   NA        NA

Any idea what could be causing this issue as well?

hanchenphd commented

Can you send me a reproducible example? My email address can be found on the GMMAT home page.

Thanks,
Han

dylkot commented

Done. Thank you!

hanchenphd commented

Thank you for sharing the example! I think the problem was that your formula was a class character object, instead of a formula:

> class(formula)
[1] "character"

This can be fixed using as.formula:

> formula<-as.formula(formula)
> class(formula)
[1] "formula"

It fixed the problem in the toy dataset that you sent to me.

Re: Issue with the GDS file: if you grep "variant.id" using seqGetData, it will return an integer vector (variant indices). To get the SNP names you would want to use "annotation/id" instead.

Best,
Han

dylkot commented

Thank you! indeed this fixes it.

Cheers,
Dylan