Tagged-Amplicon Sequencing (TAm-Seq)
TAm-Seq is the first method to demonstrate reliable de-novo identification of rare cancer mutations, by screening entire genes in degraded or dilute samples such as circulating DNA or FFPE specimens. It shows the potential for using circulating tumour DNA as a "liquid biopsy" to identify noninvasiely genomic changes in a cancer, and also provides a generalizable and cost-effective method to track tumour burden over time by a blood test. Images below are taken from our recent paper, Forshew et al., Science Translational Medicine 4, 136ra68 (2012).
TAm-Seq allows identification of rare mutations in degraded or dilute samples such as circulating DNA or FFPE. Together with colleagues in academia and industry, we developed a protocol to amplify entire genes by tiling short amplicons (left panel) through a two-step amplification process (2nd panel). This generates amplicon libraries that are tagged by sample-specific barcodes and are ready for sequencing (3rd panel). By studying the pattern of background allele fractions of non-reference base alterations across multiple samples (right panel) we designed a strategy and algorithms that reliably call rare mutations that are present above background rates in duplicate amplification of the same initial template.
The allele frequencies measured by TAm-Seq are highly reproducible (left panel) and agree with expected frequencies (2nd panel) and direct measurements by digital PCR (3rd panel). The accuracy is limited by sampling noise which is the random distribution of rare alleles (right panel).
We showed that TAm-Seq accurately identifies mutations that are present at allele frequency >2% in a sample (left panel). Together with Christine Parkinson and James Brenton's group, we showed that this allows noninvasive identification of TP53 mutations in plasma samples for the majority of patients with advanced high-grade serous ovarian cancer cancer. We also showed that TAm-Seq could be used to track changes to the frequency of the mutant alleles in serial plasma samples form ovarian cancer patients, without need for any patient-specific adaptations or assay design (2nd and 3rd panels). Together with Sarah-Jane Dawson and Carlos Caldas' group we showed that TAm-Seq could be easily adapted to monitor multiple mutations in parallel on the sample sample, providing infomation on tumour evolution and reducing risk of sampling errors (right panel). These examples showed that changes to the fraction of mutant alleles in the plasma over time reflected clinical response patterns, as demonstrated by Diehl, Diaz, Vogelstein and co-workers.