The heterogeneity of DNA variants both within and among tumors from different individuals serves as a formidable barrier to current molecularly targeted cancer therapies, and for most solid tumors the assessment of clinical response to treatment regimens may occur much later than the loss of therapeutic efficacy or the emergence of resistance. Additionally, for the majority of cancers there are no assays to detect minimal residual disease after definitive therapy. Thus, improved biomarkers are needed for the early and non-invasive assessment of molecularly targetable variants, treatment response, residual disease, tumor progression, and resistance.
As our understanding of the molecular architecture and functional complexity of cancer expands, it is evident that we need extremely sensitive and precise techniques to detect, target, and manage cancer. The techniques should be sensitive enough to detect a few mutant molecules, versatile enough to identify different classes of mutations, and powerful enough to quantify specific, relevant variants. One widely-published technique from the scientific literature, CAPP-Seq, demonstrates the ultrasensitive detection of a wide array of circulating tumor DNA (ctDNA) species from a standard blood sample.1-6 This technique combines the high throughput, sensitivity, and versatility of next generation sequencing (NGS) with adaptive informatics algorithms to precisely identify and quantify mutations in cell-free tumor-derived DNA (ctDNA).1,2
CAPP-Seq is an ultrasensitive technique that detects and quantifies mutations in ctDNA. ctDNA isolated from the blood of cancer patients is known to harbor identical genetic mutations (including point mutations, chromosomal rearrangements, amplifications and aneuploidy) to those in tumor cells.7 Liquid biopsy, a method for collecting ctDNA from blood samples to identify and quantify mutations in the tumor, has recently emerged as an alternative to tissue biopsy. This noninvasive method is highly sensitive, cost-effective, and could facilitate early detection and more frequent monitoring of cancer due the ease with which samples can be obtained from a simple blood draw. However, liquid biopsy has its own limitations. The amount of ctDNA that can be collected from plasma is very limited, the prevalence of tumor-derived, actionable variants amounts to only a handful of molecules with each mutation, and there are patient-specific differences in the pattern of tumor-derived mutations that preclude a one-size-fits-all approach. Knowing which mutations to target, targeting molecules with those mutations effectively from among the billions of molecules in the sample, and quantifying each mutation precisely is extremely challenging. CAPP-Seq provides a robust solution to all of these issues and offers users the flexibility to balance performance and sequencing cost.
CAPP-Seq uses hybridization capture based target enrichment combined with deep sequencing for maximum recovery of low input DNA, and adaptive informatics algorithms to target recurrent mutations from publicly available cancer sequencing data.1 In this method, a selector consisting of biotinylated DNA oligonucleotides targeting recurrent mutations is designed and applied to the tumor DNA to identify mutations specific to an individual. The population-level bioinformatics analysis builds the CAPP-Seq selector library, which can be used at the patient-level analysis to detect personalized biomarkers.
The CAPP-Seq selector library is constructed by iteratively adding markers specific for an individual in multiple phases. The selector is constructed with first capturing the genomic regions harboring known and suspected driver mutations and adding exons containing recurrent mutations using whole exome sequencing data in the next iteration. Exons of driver mutations as well as those containing fusion breakpoints in rearrangements of specific genes are also added in the selector subsequently to facilitate the detection of structural rearrangements. The CAPP-Seq selector is also applied to the control group with no somatic mutations (without the tumor) for comparison. Deep NGS sequencing with the selector after optimizing library prep and target enrichments steps is then performed in both groups for identification and quantification of mutations.
To minimize nonbiological errors introduced during library preparation and to increase efficient recovery of cell-free DNA during deep sequencing, an advanced bioinformatics approach called integrated digital error suppression (iDES) is used.3 This method eliminates stereotypic background artifacts by characterizing normally occurring background errors during capture-based deep sequencing using a computational approach for error suppression and background polishing. This, combined with molecular barcoding used to tag individual DNA molecules with unique identifiers, can further increase the sensitivity of cancer profiling using CAPP-Seq.