Target Enrichment for NGS

Molecular Diagnostics

Target enrichment is a general term used to describe strategies developed to selectively isolate specific genomic regions of interest for more in-depth analysis.

Target enrichment was originally developed as a lower-cost alternative to whole genome sequencing (WGS). As technological advances have rendered that approach more routine, one might question the utility of target enrichment in today’s landscape. After all, why would one spend extra time and money to enrich samples for specific regions, when access to the entire genome can now often be obtained for a fraction of its cost a few years ago?

The answer is quite simple-as scientists dig deeper into the genomic underpinnings of disease, discovery of causal mutations present at low frequencies requires increased depth of coverage for detection. This is especially true in cancer, and is further exacerbated with the heterogeneity found in solid tumors and circulating cell-free DNA. For example, where the standard offering for WGS typically covers the genome at a mean depth of about 30X, many panel-based oncology assays that utilize target enrichment will provide coverage of the tumor well exceeding 500X mean target coverage-which is still cost-prohibitive for WGS. Furthermore, clinical applications raise the question of incidental findings and how to report them, creating challenges for diagnostic assays based on sequencing entire genomes.

This article will address how to select the appropriate experimental strategies, specifically in differentiating three enrichment strategies: multiplex PCR, hybridization and other novel approaches.

Strategy Considerations

With various strategies to choose from, there are several aspects to consider when choosing a target enrichment technology. Perhaps the primary factor is the utility of the assay, with a clear distinction between a goal of identifying nucleic acid variants of well-characterized clinical utility versus a goal of aiming to discover novel nucleic acid variants that may be associated with disease. The former lends itself to more focused enrichment, while the latter is driven by tradeoffs between sequencing costs and target territory, as well as sample cohort sizes for a given study. Other aspects that have significant financial impact include overall specificity of the approach and the uniformity of coverage across targets, as these can significantly impact the cost of sequencing required to obtain the necessary sequence coverage across targets for reliable variant calling.

The type of nucleic acid variant being studied is also of critical importance. Different technologies are better suited to reliably call different types of variants, including single nucleotide variants, insertions, deletions, copy number variants and specific breakpoints associated with structural rearrangements.

Another key factor to consider is the nature of the sample being studied in terms of both the available amount and the condition of the material, particularly with regard to purity, degradation, molecule size and presence of single-stranded DNA that may be present. These will all impact the overall sensitivity of the assay, which is typically defined as how well the assay can reliably detect variants of low frequency using a given amount of input material.

Other considerations include the workflow turnaround time, reproducibility of the assay and workflow amenability to laboratory automation platforms.

Multiplex PCR Approaches

A range of multiplex PCR-based approaches are available for enriching relatively smaller numbers of targets for panels ranging from 1-100 genes, with target areas in the 10-500 kb range. These are characterized by the utilization of primer pairs to amplify multiple targets in parallel. The primers are typically tailed with sequences to add sequencing adaptors, which results in a fast workflow that utilizes low sample amounts as input material. Commercial products that fall into this category include Swift Biosciences Accel-Amplicon, Ion Torrent AmpliSeq and Archer VariantPlex.

Hybridization-Based Approaches

Hybridization-based approaches begin with production of a sequencing library that is then hybridized to long, biotinylated oligonucleotides specific to the target of interest. Hybridized molecules are separated using streptavidin beads. This strategy has been employed to enrich large numbers of targets, ranging from panels in the megabase (Mb) size range enriching hundreds of genes, to the whole exome, representing all of the protein coding regions of the genome, ranging to 60-100 Mb. Commercial products include Agilent SureSelect, Roche NimbleGen SeqCap EZ, IDT xGen Lockdown Probes and Illumina Nextera Rapid Capture.

Novel Approaches

Newer approaches have been developed in an attempt to balance the two strategies, with the goal of enabling more specific enrichment for panel sizes ranging between smaller gene content and more comprehensive panels comprised of hundreds of genes. These strategies have employed a wide range of innovative techniques to overcome inherent limitations. These tend to perform well across the range of content, yet also require complex steps that can involve splitting sample material across multiple reactions, and use of highly customized labware to achieve high specificity. Examples of these include Agilent HaloPlex, Illumina TruSeq Amplicon, and, most recently, New England Biolabs NEBNext Direct.

Tradeoffs

Hybridization-based panels work well for whole exome sequencing, where specificity is maintained and the ability to filter PCR duplicate reads is based on upfront random shearing of DNA molecules. As these technologies scale down to panels comprised of hundreds of genes or fewer, specificity drops significantly depending on the specific content. This approach also tends to have longer workflows with overnight hybridization steps and the need to create an upfront library.

In contrast, multiplex PCR-based panels show high specificity for smaller sized gene panels. Filtering of PCR duplicate molecules is difficult, however, since the nature of designing primers creates uniform start sites across input molecules, which creates difficulty for accurate assessment of variant allele frequencies. These technologies are also limited in the ability to scale content, and coverage uniformity tends to be poor based on the need to balance enrichment of specific amplicons.

Overall, the wide range of approaches makes choosing the right target ðenrichment technology for a given application difficult, and a one-size-fits-all approach will likely represent tradeoffs in scientific utility. Finally, as the boundaries of what can be discovered and detected expands, scientists and reagent technology providers must continue to innovate in this space to create new techniques capable of answering fundamental biological questions related to the genomic basis of disease.

Andrew Barry is product marketing manager, Targeted Enrichment, New England Biolabs, Inc.