EXCAVATOR workflow. BAM files of both test and control samples are processed by means of SAMtools and R scripts for EMRC calculations. After EMRC calculation, EXCAVATOR corrects the data for GC-content, mappability and exon size. After normalization, normalized read count (NRC) for each sample are organized according to the analysis mode (pooling or somatic) selected by the user: pooling mode to compare one sample to a pool of normal controls, somatic mode to compare one sample to its corresponding normal control. Finally, HLSM and FastCall are applied to normalized data and results are provided as tab-delimited text files (variant call format, VCF and BED format). HSLM, heterogeneous shifting level model; RC, read count.