The GATK callers (HaplotypeCaller and UnifiedGenotyper) are by design very lenient in calling variants in order to achieve a high degree of sensitivity. This is a good thing because it minimizes the chance of missing real variants, but it does mean that we need to refine the call set to reduce the amount of false positives, which can be quite large. The best way to perform this refinement is to use variant quality score recalibration (VQSR). In the first step of this two-step process, the program uses machine learning methods to assign a well-calibrated probability to each variant call in a raw call set. We can then use this variant quality score in the second step to filter the raw call set, thus producing a subset of calls with our desired level of quality, fine-tuned to balance specificity and sensitivity.
The downside of how variant recalibration works is that the algorithm requires high-quality sets of known variants to use as training and truth resources, which for many organisms are not yet available. It also requires quite a lot of data in order to learn the profiles of good vs. bad variants, so it can be difficult or even impossible to use on small datasets that involve only one or a few samples, or on targeted sequencing data. If for either of these reasons you find that you cannot perform variant recalibration on your data, we recommend you use hard-filtering instead. See the methods articles and FAQs for more details on how to do this.