Contributed by Dr. Samuel Levy : Director of Genomic Sciences - Scripps Translational Science Institute and Scripps Genomic Medicine, Professor of Molecular and Experimental Medicine - The Scripps Research Institute

AmpliSeq Cancer Panel applied to a tumor and blood sample pair for the detection of common somatic mutations

We have a clinical research protocol that aims to provide tumor mutation data to oncologists as part of a patient’s cancer care within the Scripps network and we were trying to figure out ways of providing these data in a time sensitive manner to patients and their physicians.  Whilst both exome and whole genome sequencing are good options, we wanted to maximize the use of any large amount of data we generate in an oncologist’s care approach. Consequently we decided that these options were data overkill for what a treating oncologist typically employs.

For this particular application we turned to the Ion Torrent and the AmpliSeq Cancer panel that provides a rapid manner to assess the presence of a set of mutations in a DNA sample. EdgeBio were able to generate these sequence data to us and we tested this out on a colon tumor sample and the germline DNA from white blood cells from the same patient.  We sequenced 190 amplicons from a single multiplex reaction targeting 46 genes contain 739 known SNVs or indels identified from a variety of cancer samples (Life Technologies AmpliSeq Cancer Panel). Using the Ion Torrent 316 chips we were able to obtain >~5,000 fold average coverage of the 13.3 kb of target sequence from the tumor and normal sample pair.  The uniformity of sequence coverage for the sample pair was high (Figure 1), enabling variant calls to be made in a consistent and comparable manner.

Figure 1. Normalized coverage plot for tumor and normal sample pair. The average coverage across the target regions was normalized to unity enabling comparisons between samples.  For example, 80% of each sample was covered at 2,000x (0.4 x 5,000). Note that all target regions have at least 10x coverage and no PCR dropouts were detected.

High coverage and appropriately applied variant calling filters provided in the Ion Suite software enabled the identification of somatic mutations in PIK3CA (Q546K - 38% mutation frequency), KRAS (G13D – 54% mutation frequency) and TP53 (R273L – 75% mutation frequency). We verified that these changes do not occur in the blood sample as evidenced by the comparative data from read counting in Figure 2. The data we identified here were fully consistent with the existing data we had on this sample using other sequencing approaches (Illumina – whole genome and whole exome sequencing) and provides a useful and rapidly generated mutational profile on a tumor sample for clinical applications.

Figure 2. Read alignment of tumor sample versus blood in the KRAS gene depicting the G13D mutation. Reads containing the mutated allele have the base “T” depicted in red.