There are many filtering rules, but I would like to draw your attention to the need for a reasonably good coverage (clearly above 10x). Refer to Meynet et al., Variant detection sensitivity and biases in whole genome and exome sequencing, at http://www.biomedcentral.com/1471-2105/15/247/.
bcbio-nextgen is a Python-based pipeline for high-throughput sequencing analysis, which is worth checking out (https://bcbio-nextgen.readthedocs.org).