Hitting the Mark with Targeted Sequencing

There can be no doubt genome sequencing has irrevocably altered the practice of both research and medicine. Scientists routinely turn to the genome to identify interesting genetic associations and variants, and clinicians increasingly are doing likewise to make diagnoses that cannot be determined any other way. Even drug development is changing, as evidenced by AstraZeneca’s late April announcement of plans to sequence some two million human genomes over the next decade in a bid to identify interesting genetic variants, and the drugs that potentially can target them.

These developments stem from the plummeting cost and exploding availability of next-generation DNA sequencing (NGS). Still, not every researcher opts for whole-genome analyses, whether for economic, bioinformatic or other considerations. For many such scientists, targeted sequencing is all they need.

The benefits of targeting

Targeted sequencing is precisely what it sounds like: sequencing only that portion of the genome that is actually of interest, whether a single gene region, a multigene panel or a complete exome. Swift Biosciences’ Accel-Amplicon™ Comprehensive TP53 panel focuses on 21 regions of the p53 tumor suppressor, for instance, while Roche’s SeqCap EZ MedExome Enrichment Kit targets 47 Mb of content, including the complete exome with particularly targeted coverage of regions deemed medically relevant by an academic consortium.

Whatever its scale, targeted sequencing offers several key benefits, says Andy Felton, vice president of product management for next-gen sequencing and oncology products at Thermo Fisher Scientific. For one thing, targeted sequencing enables researchers to reduce per-sample costs (and/or boost per-sample sequencing depth) by squeezing more samples into each sequencer run.

Genomes typically are sequenced to 30x coverage or so, for instance. But to find rare variants, researchers need to read much deeper, to 1000x or higher, an impractical goal when sequencing every base in the genome.

On the computational side, bioinformatic complexity and data-storage requirements decrease with more targeted datasets.

And then there’s the matter of what the data mean. In oncology, for instance, researchers and clinicians tend to be interested in a relatively limited set of actionable and interpretable markers, Felton explains—genes whose mutational status can inform treatment decisions. “You can’t do anything with the extra information”—noncoding sequences and genes of unknown significance, for instance—“at all.”

Available options

Researchers basically have two choices when it comes to enriching regions for targeted sequencing. In solution hybridization-based methods, genomic DNA is hybridized with single-stranded, biotinylated oligonucleotides complementary to the regions of interest, which are then pulled down using streptavidin-coated beads. Multiplexed PCR-based methods use multiple sets of primer pairs to selectively amplify regions of interest in a single reaction.

Examples of the former approach include Agilent Technologies’ SureSelect, Roche Diagnostics’ SeqCap EZ and IDT’s xGen® Lockdown® panels. Agilent’s new OneSeq system, based on SureSelect, targets 28 Mb of genomic DNA and “consists of an optimized genome-wide backbone plus a mutation panel that allows one to confidently identify copy number variations, mutations and loss of heterozygosity using NGS,” says Rebecca Brandes, director of product marketing for Agilent Technologies’ Diagnostics and Genomics Group.

PCR-based systems include Thermo Fisher Scientific’s Ion AmpliSeq™, Agilent’s HaloPlexHS and Swift Biosciences’ Accel-Amplicon panels. Accel-Amplicon’s single-tube workflow “enables a rapid, two-hour prep with the ability to detect 1% allele frequencies from 10 ng of input,” says Laurie Kurihara, Swift’s director of research and development.

Illumina’s TruSight™ reagents are based on both technologies; the company’s TruSight One Sequencing Panel uses hybridization to target more than 4,800 “clinically relevant genes,” and its TruSight Myeloid Sequencing Panel uses PCR to target 54 genes frequently mutated in leukemia.

Though either strategy will work in most situations, hybridization-based methods generally are used to pull down relatively large targets, such as whole exomes or large multigene panels, while.PCR-based methods tend to be reserved for smaller panels.

“PCR-based approaches, like HaloPlex, tend to have simpler workflows and are more amenable to smaller targeted regions. Capture-based approaches like SureSelect tend to involve a couple more steps but offer higher workflow flexibility and capture larger design regions,” Brandes says.

There are, of course, exceptions: The Thermo Scientific Ion AmpliSeq Exome Panel comprises 239,903 amplicons in a dozen pools of nearly 26,000 amplicons each, for instance, and New England Biolabs’ (NEB’s) new NEBNext Direct™ Cancer Hotspot Panel, based on hybridization, targets 37 kb of sequence across 190 sites and 50 genes. According to Andrew Barry, target enrichment product market manager at NEB, the NEBNext Direct system takes just seven hours “from purified genomic DNA to a sequence-ready library.”

Target selection

According to Felton, two key metrics to consider in selecting a target-capture approach are on-target rate and uniformity.

Uniformity, he says, defines the percentage of amplicons (or targeted regions) that are “within some range of targeted median coverage,” and on-target rate refers to the number of reads that correspond to the targeted regions. Both impact the amount of sequencing required, as off-target reads are basically wasted bases, while uneven coverage—10x coverage on some regions and 2000x on others—inevitably requires deeper sequencing than might otherwise be necessary.

Another key metric is the required DNA input—a key consideration when working with precious samples, such as cell-free DNA, single cells and formalin-fixed paraffin-embedded (FFPE) material. “Building libraries from CTCs [circulating tumor cells], FFPE and cell-free DNA is really taking off for us,” says Kurihara. Indeed, Swift Biosciences’ Accel-NGS 2S Hyb library prep kit, optimized for low input levels, “enables exome sequencing from 1 ng input DNA,” she says. AmpliSeq kits require as little as 10 ng of DNA, according to product specifications, compared with 50 ng for Agilent’s SureSelectQXT reagents and 100 ng for the SeqCap EZ MedExome.

And finally, there’s customizability. Most companies offer pre-designed human exomes and/or select panels targeting popular research foci, such as oncology or neurobiology. Users working with nonstandard model organisms or pathways, however, may need to generate their own target designs, either by building them from scratch or by adding custom content to and selecting subsets of existing panels, says Mike Leous, group marketing manager for sequencing products at Roche Diagnostics.

Roche customers can either design panels themselves, using the company’s online NimbleDesign tool, or work with one of the company’s staff bioinformaticians, Leous says, targeting up to 200 Mb total using Roche’s SeqCap EZ Choice XL enrichment kit.

Similar tools, options and services are available from other vendors, as well, including Agilent, Thermo Fisher Scientific, IDT and Illumina.

Targeted sequencing is an effective alternative to whole-genome sequencing and is enabling researchers to obtain a more detailed and focused view of specific regions of interest in the genome. Tool providers continue to improve the reagents and develop custom products, enabling researchers to concentrate on interpreting results and building understanding of complex biomarkers and their regulatory responses in the genome.