Experimental variables affecting data quality and accuracy for detecting mutations. (A) Panel-based, WES and WGS are compared in their completeness and requirements for volume of raw data (reads). Each method can be applied to tumors in isolation or with matched germ line DNA also sequenced (T/N pairs). Only common single nucleotide polymorphisms can be recognized from variants detected in unmatched tumors, whereas rare germ line variants and private mutations can be removed if the germ line is sequenced. (B) Frozen tissue is a preferred source of genomic DNA. FFPE-derived DNA has various forms of DNA damage and can be highly fragmented, leading to more overlapping read pairs and more redundant sequence information from each fragment. (C) It has been common to report the average total depth of depth regardless of fragment length or read overlap. Effective (or corrected) depth accounts for the redundant information from overlapping reads and is more consistent with how variant calling algorithms detect mutations.