* Fixed the bug where "CD4" in the output VCF file where alternate concordant reads where grabbed twice, when it should've been alternate concordant and then alternate discordant read.
* Added (limited) tumor-only support.
* Convert VarDict's "Complex" variants into SNVs when appropriate.
* Slightly modified r_scripts/ada_model_builder_ntChange.R script, i.e., the arguments succeeding the input TSV file are features to be ignored in training.