I would like to convert possibly-uncertain genotypes from a VCF to a snpMatrix object. My VCF file was produced by GATK 3.4-46 and complies with VCF specification 4.1. Here is what I am doing:

vcf <- readVcf(file=tabix.file, genome="test", param=vcf.params)
snp.mat <- genotypeToSnpMatrix(vcf, uncertain=TRUE)
In .local(x, ...) : uncertain=TRUE requires GP or GL; returning NULL

Could it be possible to add support for PL (phred-scaled genotype likelihood) ?

Here is my sessionInfo():

R version 3.2.2 (2015-08-14)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS release 6.6 (Final)
attached base packages:
[1] stats4 parallel stats graphics grDevices utils datasets
[8] methods base
other attached packages:
[1] snpStats_1.20.0 Matrix_1.2-2
[3] survival_2.38-3 VariantAnnotation_1.16.0
[5] Rsamtools_1.22.0 Biostrings_2.38.0
[7] XVector_0.10.0 SummarizedExperiment_1.0.0
[9] Biobase_2.30.0 GenomicRanges_1.22.0
[11] GenomeInfoDb_1.6.0 IRanges_2.4.0
[13] S4Vectors_0.8.0 BiocGenerics_0.16.0

VariantAnnotation 1.17.8 adds support for the PL field. It should be available later today after propagating through the build process.