GATK not calling varients found using samtools
I'm trying to compare my custom SNP-calling pipeline to GATK, and have found that GATK doesn't seem to be calling some SNPs that my pipeline found. For my pipeline, I'm looking at proportion of samples with each identity at a position (requirement of a depth of at least 1000 in all samples at the position before attempting calling), as determined by a file from samtools mpileup (with depth adjusted to avoid downsampling), and calling a SNP if they aren't all the same. For GATK, I'm using HaplotypeCaller with no downsampling, --forceActive, and a ploidy of 1 (with the same 18 samples in each case). My question is: why might GATK not find some SNPs that my pipeline finds?