ValueError: could not convert string to float: '.'

[algarnham]

Hi,

I've been trying to run vireo (v0.5.8) for some snRNAseq data, but I've been getting the following error:

Loading vireoSNP/0.5.8
Loading requirement: bcl2fastq/2.19.1
[vireo] Loading cell folder ...
[vireo] Loading donor VCF file ...
[vireo] 3310 out 3470 variants matched to donor VCF
Traceback (most recent call last):
File "/stornext/System/data/apps/vireoSNP/vireoSNP-0.5.8/bin/vireo", line 11, in
load_entry_point('vireoSNP==0.5.8', 'console_scripts', 'vireo')()
File "/stornext/System/data/apps/vireoSNP/vireoSNP-0.5.8/lib64/python3.6/site-packages/vireoSNP/vireo.py", line 165, in main
options.geno_tag)
File "/stornext/System/data/apps/vireoSNP/vireoSNP-0.5.8/lib64/python3.6/site-packages/vireoSNP/utils/vcf_utils.py", line 306, in parse_donor_GPb
GT_prob[i, j, :] = parse_GT_code(GT_dat[i][j], tag)
File "/stornext/System/data/apps/vireoSNP/vireoSNP-0.5.8/lib64/python3.6/site-packages/vireoSNP/utils/vcf_utils.py", line 288, in parse_GT_code
_prob[int(float(code[0]) + float(code[-1]))] = 1
ValueError: could not convert string to float: '.'

The command I'm using is:

vireo --cellData cellsnp_lite --donorFile all_donors_chr.vcf.gz --outDir vireo_strat2 --nInit 200 --nproc 30 --genoTag GT

The all_donors_chr.vcf.gz has been filtered using the cellsnp_lite regions as recommended in the manual.

Any advice/help you could give in over coming this error would be much appreciated!

Thanks!

[huangyh09]

Hi, thanks for raising the issue and sorry for the delay.

It looks like in your donorFile all_donors_chr.vcf.gz, there are SNPs with uncommon genotypes (usually, ., .|. or ./.).
https://github.com/single-cell-genetics/vireo/blob/master/vireoSNP/utils/vcf_utils.py#L284

You may check if you only include biallelic SNPs. Otherwise, please only keep bi-allelic SNPs.

Yuanhua