Filtering variants in RNA-seq data

[chadn737]

I know that when making variant calls from genomic samples its common to set a max read depth. However in RNA-seq data, I have known variants (confirmed by sanger sequencing) in highly expressed genes which get filtered out because of the high coverage.

So I know using typical filters, I am missing real variants. What settings, especially for read depth, are optimal for calling variants from RNA-seq data?