Locus Zoom Issuing "No valid markers in VCF file" on a VCF created in Plink 1.9

Question

[LauferVA]

What follows is output from locus zoom followed by a sample of the first 20 lines and first 12 columns of a vcf file. The vcf file is bgzipped and tabixed. It was created directly from plink1.9.

I'm receiving the message "Error: no valid markers in VCF file to compute LD from.."

I am writing to ask if anyone has encountered a similar problem while using either locuszoom or plink1.9. I have not contacted the authors of locuszoom because in the past I never received a response to another query. Any help would be much appreciated.

LZ output:

Using /share/apps/ngs-ccts/locuszoom/locuszoom-1.3/bin/locuszoom.R..

Beginning plotting sequence for: chr5_7512611-8512611

Extracting region of interest (chr5:7512611-8512611) from input file..

Attempting to find best SNP in region..

Found: chr5:8012611

Using user-specified VCF file to calculate LD with reference SNP chr5:8012611..

Error: no valid markers in VCF file to compute LD from..

Grabbing annotations from SQLite database..

Creating plot..

Warning message:

No usable LD information.

Deleting temporary files..

Time required: 0d:0h:0m:35s

command used to display file: zcat MTRR_area_ra_susc.vcf.gz | head -20 | awk '{print $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12}'

output of command ##fileformat=VCFv4.2
##fileDate=20160531
##source=PLINKv1.90
##contig=<id=5,length=9999893>
##INFO=<id=pr,number=0,type=flag,description=provisional reference="" allele,="" may="" not="" be="" based="" on="" real="" reference="" genome=""> ##FORMAT=<id=gt,number=1,type=string,description=genotype>
#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT 0_1501_1 1_1502_1 2_1503_1

5 6000051 rs17818165 C T . . PR GT 0/1 0/0 0/0

5 6000064 rs466640 A T . . PR GT 0/1 0/0 0/1

5 6000282 rs7718138 C T . . PR GT 0/0 0/0 0/0

5 6000403 rs823742 T C . . PR GT 0/1 0/0 0/1

5 6000494 rs823743 C T . . PR GT 0/1 0/0 0/1

5 6000584 rs258972 G A . . PR GT 0/0 0/0 0/1

5 6000847 rs823568 C T . . PR GT 0/1 0/0 0/1

5 6001581 rs466015 A G . . PR GT 0/1 0/0 0/1

5 6001813 rs460967 C G . . PR GT 0/1 0/0 0/1

5 6001868 rs461940 A G . . PR GT 0/1 0/0 0/1

5 6001992 rs464210 A T . . PR GT 0/1 0/0 0/1

5 6002069 rs435787 G T . . PR GT 0/1 1/1 0/1

5 6002085 rs258971 C G . . PR GT 0/0 0/0 0/0

The actual VCF file is tab-delimited, the quick and dirty awk command is why it becomes single space delimited. Has anyone dealt with an issue like this?

Answer 1

[chrchang523]

First thing I'd try is converting the spaces back to tabs, since the VCF format is explicitly defined to be tab-delimited. "cat [space-delimited VCF filename] | tr ' ' '\t' > [new VCF filename]" should do the trick.

[LauferVA]

Hi Chris!

Thanks very much for your reply.

The VCF file is tab-delimited. The post I posted here is space delimited, due to the awk line that I used to display it here for people to look at. What I was trying to say is that the single-spacing is not an issue, because the VCF file is correctly formatted, despite the way I reprinted here.

Sorry for any confusion that caused.

What would you try next?

[chrchang523]

The LocusZoom documentation states that a tabix index is required for the VCF file; do you have that?

Next thing I'd try after that is phasing the calls.

[LauferVA]

yes, the file is tabix indexed. why would phasing be necessary if the metric being used is r^2 and not D prime?

Answer 2

[clisun]

Hello, Laufer I have the same issue. Did you solve the problem Thanks Charlie

[chrchang523]

If you type "plink --version" for the version of plink that generated the VCF, what do you get?

Answer 3

[pjawinski]

I noticed that plink creates .vcf files with a single '.' for each variant in the FILTER column (as shown above). However, locuszoom's vcf_ld.py function checks whether variants passed filters as indicated by 'PASS' in the FILTER column. A simple solution is to call locuszoom with the flag '--ignore-vcf-filter'. Best, Philippe