Failing to produce calls for same samples

[buchanri]

Using graphtyper to create a vcf we are running into a problem of the vcf is not producing calls for samples that do have variants. Double checking with IGV we can see that the bam file of specific samples does have variants but when compared with the vcf, the vcf doesn't show that there are variants. We checked the average read depth of the samples using samtools and they do have good read depths.
Here's a screenshot of IGV comparison to the vcf (top row) and a sample (Kew122) that does have variants.

Graphtyper did work for most of our samples which looks like: