Cancer samples, and biopsy in particular, can generate low amounts of total DNA. Whole genome amplification (WGA) by multiple strand displacement is a popular method to increase the amount of material available for clinical assays [20]. We evaluated the effect of WGA on the CAL-B calibration sample by comparison with the non-amplified sample. Both sensitivity and PPV were unchanged (Figure S7 in Additional file 1). Surprisingly, in the WGA-amplified samples the observed prevalence of mutations expected at 5% or less dropped significantly (Figure 1f). This is likely due to allele-specific bias generated during the amplification. We then applied WGA to the breast cancer xenograft sample and performed a UDT-Seq assay, observing the HRAS-G12V mutation at 49%, in agreement with the prevalence observed without WGA (Table 1). Because the initial sample did not carry any low prevalence mutations, we could not verify the potential allelic bias below 5%. Thus, UDT-Seq analysis of DNA samples subjected to WGA provides reliable results for highly prevalent mutations but underestimates the presence of low prevalence alleles.
