Advanced sequencing platforms are gradually edging toward clinical applications
By Steve Halasey
Sometime in the not-too-distant future, say the experts, clinical laboratories will gradually begin to adopt next-generation sequencing (NGS) among the standard methods used to identify pathogens and diagnose disease.
But such a shift won’t happen overnight. And it won’t happen next week. Just when NGS will be ready for prime time in clinical settings remains something of a puzzle, as even the experts are uncertain about which of many obstacles may prove to be the one that slows the march toward adoption—or even stops it in its tracks.
Taking stock of the progress of NGS toward clinical adoption is the subject of a new report compiled by the Association for Molecular Pathology (AMP) and published online at the end of September.1 Focusing exclusively on the application of NGS to infectious disease testing, the report devotes special attention to genotypic resistance testing, direct detection of unknown pathogens in clinical specimens, microbial population diversity in human hosts, and strain typing.
“This article is a timely, clearly written, and comprehensive review of the current state of NGS, but also of future applications for NGS in the diagnosis and monitoring of infectious diseases,” says Michael A. Lewinski, PhD, D(ABMM), CM, MB(ASCP)CM, senior director of clinical science for microbiology and CT/NG at Roche Molecular Systems Inc, Pleasanton, Calif, and chair of AMP’s infectious diseases subdivision. “In addition, the article discusses the technical and bioinformatics challenges that should be addressed as this technology advances toward clinical practice.”
Authors of the AMP report acknowledge that NGS still has a steep hill to climb on the way to becoming a mainstay in clinical laboratories. “Although NGS holds enormous promise for clinical infectious disease testing, many challenges remain, including automation, standardizing technical protocols and bioinformatics pipelines, improving reference databases, establishing proficiency testing and quality control measures, and reducing cost and turnaround time—all of which would be necessary for widespread adoption of NGS in clinical microbiology laboratories,” they write.
In some cases, the obstacles that lie ahead represent a series of interrelated issues that researchers may need to address at the same time. “Clearly, turnaround time can be improved with the adoption of automation. When specimen processing and library creation become automated, that will certainly help to reduce turnaround time,” says Lewinksi.
“In the beginning, as NGS testing is adopted for more and more clinical applications, the measure of clinical utility may focus on reducing turnaround times so that the testing is seen as clinically relevant and actionable,” adds Lewinksi. “But we will need to be careful and define what we mean by clinically actionable.
“What’s clinically actionable is actually a very profound issue. In my view, emergent situations call for a stat response that includes point-of-impact testing—that is, testing specifically designed to provide actionable results within a therapeutically meaningful timeframe,” says Lewinski. “But every institution defines what it means to have a stat response somewhat differently, and varied medical conditions may also suggest different interpretations of when a test result is considered clinically actionable. I doubt that NGS will ever be actionable at the level of a 15-minute result. That’s not going to happen.”
The authors of the AMP report also identify the need to have a significant database of reference genomes as a known limitation on the utility of NGS testing. As new applications are developed, this is an issue that will need to be addressed.
“The pace at which such databases will evolve seems likely to parallel clinical need and use. One feeds the other,” says Lewinski. “Unfortunately, this means that progress may be uneven, and may appear in the form of a series of analyte-specific advances. As testing for a particular pathogen or disease reveals the need and the opportunity, databases will be developed and managed. Already, there are a number of 16S ribosomal sequence databases that everybody uses, and they need to be continually updated and validated.”
RUO, OR NOT?
In spite of the extensive development and promotional activities being undertaken by NGS companies, the regulatory status of NGS tests remains murky. A growing number of NGS assays are available for use on NGS instruments, but they are almost universally designated for research use only.
“For true clinical adoption, a product would need to be cleared by FDA. In the case of an in vitro diagnostic, that’s a two-part process requiring clearance for both the instrument and the assay,” says Lewinski. “But so far, there is only one FDA-cleared NGS instrument—the MiSeqDx by Illumina—and just two FDA-cleared assays—Illumina’s own cystic fibrosis assays that run on the MiSeqDx.
“Most of the NGS tests that labs are running right now are laboratory-developed procedures performed in a laboratory environment in compliance with the Clinical Laboratory Improvement Amendments of 1988 (CLIA)” he adds. “And that’s where the search to demonstrate the clinical utility of NGS begins.”
Several companies are on the market with NGS instruments that can be used in CLIA environments, notably including Pacific Biosciences, Roche Molecular Systems, and Thermo Fisher Scientific, says Lewinski. “As indicated by the AMP authors, a number of CLIA-compliant NGS assays are commercially available—for HIV tropism, for example—and they clearly meet the prerequisites to be considered for adoption.
“I assume that both the instrument manufacturers and the test developers are working toward FDA clearance for their products,” Lewinski adds. “But in the short term, until those products receive regulatory clearance, clinical adoption of NGS is going to be driven by laboratory-developed procedures performed in a CLIA-certified environment.”
As for the status of NGS testing under CLIA regulations, simplification for use by laypersons seems unlikely. “I don’t foresee NGS systems becoming CLIA-waived any time soon,” says Lewinski. “Almost certainly, these procedures will remain categorized as either moderate- or high-complexity—nonwaived is the new CLIA term—for some time into the future. If next-gen sequencing ever evolves to the point at which it could be considered for waived status, we will all be very happy. But that’s years away, if it is possible at all.”
THE BIG GORILLA
Without a doubt, Illumina, San Diego, is the dominant company in the field of next-generation sequencing. Not only does the company’s product line respond to the volume and throughput requirements of sequencing researchers in a variety of settings, the company also boasts the only sequencer so far cleared by FDA for in vitro diagnostic applications, the MiSeqDx.
Equally important, Illumina has so far managed to capture and hold a share-of-market lead against all comers. The numbers speak for themselves: at the end of 2014, Illumina had an installed base of more than 3,000 MiSeq instruments, 1,850 HiSeq instruments, and 500 NextSeq instruments. In addition, the company reported orders for more than 200 of its ultra-high-throughput (and ultra-high cost) HiSeq X systems. By way of contrast, the installed base of sequencers manufactured by Pacific Biosciences (discussed later), is around 150.
With such a strong and varied base among researchers, Illumina has been able to identify a wide range of opportunities to develop sequencing applications that grow its business even more. A straightforward example of Illumina’s strategy was provided at the end of September, when the company launched TruSight Tumor 15, an NGS panel designed to identify sequence variants in 15 genes commonly associated with marketed cancer therapeutics.
“In 2014, Illumina catalyzed the Actionable Genome Consortium (AGC), bringing together thought leaders to define the standards and content for NGS panels intended for use in oncology applications,” says Francis deSouza, president of Illumina. “The TruSight Tumor panel represents the first product, intended for research use only, based on the initial standards defined by the AGC and pharma partners, enabling researchers to advance the application of NGS in this important field.”
The new sequencing panel is optimized for damaged and degraded formalin-fixed, paraffin embedded (FFPE) tumor samples, enabling low frequency somatic variant detection from limited nucleic acid inputs. When paired with the Illumina MiSeq system, the panel delivers high-quality sequencing with key content and features needed for tumor analysis in translational research.
Illumina is committed to collaborating with pharma partners, the Actionable Genome Consortium, regulatory agencies, and key opinion leaders to establish best practices and consensus standards in NGS testing.
“Tumor biopsies are collecting ever smaller amounts of cancer tissue, and the fixation and preservation of the material results in degraded DNA. The challenge for scientists has been to get sufficient intact DNA to efficiently identify the key genetic drivers,” says John Leite, PhD, vice president of oncology at Illumina. “This new panel is designed to reduce the resource burden on small- and medium-sized labs by offering a simple, accurate, and fast solution that addresses tissue considerations on the front end and provides streamlined analytics on the back end.”
Jamie Platt, PhD, vice president of genomic solutions at Molecular Pathology Laboratory Network Inc, Maryville, Tenn, evaluated the product in beta testing and said the panel’s simplicity and ease of use makes it well-suited for translational labs. “The value and appeal of next-generation sequencing is the potential to consolidate traditionally iterative tumor analyses,” says Platt. “Our evaluation of the new TruSight Tumor 15, with its streamlined library prep and sequencing workflow, gives us confidence that this application is ideal for deployment. Even our most challenging samples produced results, including those with low nucleic acid inputs, giving us assurance we can analyze our most precious specimens.”
TruSight Tumor 15 offers a sample-to-data solution for research of common somatic variants. The panel’s focused gene content was informed by pharmaceutical partners, an independent consortia of key opinion leaders, and the needs of pharmaceutical clinical researchers to identify relevant somatic variants in common solid tumors.
The panel’s simple sample-to-data workflow is optimized for Illumina’s MiSeq instrument, and improves lab operational efficiency by replacing the cost- and resource-intensive practice of iterative single-gene analysis. The test’s streamlined multiplex PCR-based library prep method, with limited hands on time, produces high-quality data in approximately 36 hours from extracted DNA to result. The assay detects somatic mutations at 5% variant frequency and maximizes sample success with just 20 ng of DNA isolated from FFPE samples. According to the company, TruSight Tumor 15 is available for order and will begin shipping during the final quarter of 2015.
Illumina’s engagements with the research community are as strong internationally as they are in the United States. In a recent example, also from the end of September, Illumina entered into a strategic collaboration with Amoy Diagnostics Company Ltd, Xiamen, China, to accelerate the adoption of precision medicine and targeted therapies in China.
Under the collaboration, Amoy will develop and commercialize a series of oncology-related tests based on Illumina’s platforms, including the TruSight Tumor 15. The collaboration reflects the commitment of both companies to provide integrated solutions to meet clinical needs in China.
“Our genetic tests for detection of EGFR, KRAS, NRAS, and BRAF mutations, and ALK and ROS-1 fusions, have led the way for precision medicine adoption in clinical oncology across China,” says Limou Zheng, PhD, president and CEO of Amoy. “However, as the number of clinically actionable genetic variants in cancer increases, NGS technology is crucial in order for companion diagnostics to keep pace. Illumina is the world-leader in NGS and a natural partner for our company. We intend to combine Amoy’s expertise in molecular diagnostics with Illumina’s superior NGS technology to better serve Chinese cancer patients.”
“Genetic changes are a major cause of cancer development, and the number of clinically actionable genomic variants is growing rapidly,” says Richard D. Klausner, MD, senior vice president and chief medical officer of Illumina. “NGS technology, with its ability to generate and analyze large-scale data and its high sensitivity in detecting rare mutations, has shown great value and potential for oncology. Working with market leaders such as Amoy is central to Illumina’s strategy to advance personalized cancer diagnostics globally.”
Illumina may have captured the imagination of genomic researchers around the world—as well as the lion’s share of their market—but that doesn’t mean that competitors are ceding the ground. As previously unimagined avenues of research are discovered and explored, new market niches are also being opened, creating a competitive environment among manufacturers of NGS technologies.
At the beginning of September, Thermo Fisher Scientific, South San Francisco, Calif, introduced two new NGS systems based on the acknowledged capabilities of its Ion Torrent technology, acquired as part of its 2014 purchase of Life Technologies, Carlsbad, Calif.
Designed to deliver a comprehensive solution to simplify targeted sequencing, the new Ion S5 and Ion S5 XL benchtop systems provide scientists with rapid, cost-effective, and flexible platforms that can be scaled for a wide range of research areas, including inherited disorders, infectious agents, microbial identification, and translational cancer research.
“Ion Torrent technology has played a pivotal role in the success of the Saudi Human Genome Program, a national research project which to date has led to key published findings related to inherited disease research in the kingdom,” says Brian Meyer, PhD, chairman of the department of genetics research at King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia. “The simplicity and streamlined workflow of the new Ion S5 system demonstrates the next level of innovation for the platform, and it equally raises the bar for this important work in our lab.”
The Ion S5 and Ion S5 XL systems combine the ability to sequence gene panels and small genomes—as well as exomes, transcriptomes, and even custom assays—on a single platform. The platforms are designed with plug-and-play cartridge-based reagents to make setting up and operating the sequencers simple and efficient. With reduced hands-on time and streamlined workflow, the Ion S5 and Ion S5 XL systems make targeted sequencing more accessible to academic, translational, and clinical research labs.
The new systems require 15 minutes of manual set-up time per sequencing run, and less than 45 minutes of hands-on work from DNA to data when using the Ion Chef system for automated Ion AmpliSeq library construction, template preparation, and chip loading. The systems feature three new Ion chips—the Ion 520, 530, and 540 chips—which offer outputs of 5M to 80M sequencing reads generated in just 2.5 hours.
The Ion S5 system is suited for academic or translational labs in need of an all-in-one sequencer without turnaround time constraints. The Ion S5 XL system is suited for clinical research labs requiring faster turnaround or higher sample throughput per day. Both sequencers are currently designated for research use only. Thermo has not announced any plans to seek FDA clearance for clinical use of the systems.
Focusing on translational research, Meenakshi Mehrotra, PhD, and Dzifa Yawa Duose, PhD, at the molecular testing developmental laboratory in the department of translational molecular pathology at the University of Texas MD Anderson Cancer Center, compared the Ion S5 XL system with the Ion PGM and Proton.
“We found the Ion S5 XL system reduced the run time from 4.5 hours to 2.5 hours, and halved the data analysis time,” says Duose. “We ran different panels on the same chip and obtained 100% variant concordance with previously tested PGM and Proton platforms. The different chips also allowed us to run both smaller panels and larger panels on the same instrument.”
“The genomics information revolution of the past decade has ushered in an exciting new era for a broad range of industries seeking to tackle fundamental problems of health, safety, and security. We have developed the next generation of benchtop sequencing that delivers best-in-class workflow simplicity, integrated data analysis, and system economics,” says Chris Linthwaite, president of genetic sciences at Thermo Fisher Scientific. “The Ion S5 system leverages many of the inherent advantages offered by targeted sequencing, and it provides a comprehensive solution that is more cost-effective, faster, and easier to use than any other sequencer in the market.”
Thermo isn’t the only instrument manufacturer with a platform worthy of further elaboration to meet the needs of newly emerging research markets. Pacific Biosciences, is another leader in the field, and is a pioneer in long-read sequencing using its single molecule, real-time (SMRT) technology.
At the end of September, Pacific Biosciences announced the launch of a new nucleic acid sequencing platform called Sequel. Compared to the company’s previous sequencer, the PacBio RS II system, the Sequel system provides higher throughput, more scalability, a reduced footprint, and lower sequencing project costs—all while maintaining the existing benefits of the company’s SMRT technology.
The core advancement embodied in the Sequel system resides in the capacity of its redesigned SMRT cells. At launch, Sequel’s SMRT cells each contain one million zero-mode waveguides (ZMWs)—a significant increase over the cells of the PacBio RS II, which contained 150,000 ZMWs. Immobilized within the ZMWs are active individual polymerases, providing windows to observe and record DNA sequencing in real time. The Sequel system is able to perform about seven times as many reads per SMRT cell as the PacBio RS II was able to perform. According to the company, customers should be able to realize lower costs and shorter timelines for sequencing projects, with approximately half the upfront capital investment compared to the previous technology. The US list price for the Sequel system is $350,000.
Pacific Biosciences’ products enable scientists to perform a number of operations essential for resolving genetically complex problems:
- De novo genome assembly, finishing genomes in order to more fully identify, annotate, and decipher genomic structures.
- Full-length transcript analysis, to improve annotations in reference genomes, characterize alternatively spliced isoforms in important gene families, and find novel genes.
- Targeted sequencing to more comprehensively characterize genetic variations.
- DNA base modification identification to help characterize epigenetic regulation and DNA damage.
According to Pacific Biosciences, the company’s SMRT technology provides industry’s highest consensus accuracy over the longest read-lengths, in combination with the ability to detect real-time kinetic information. The Sequel system, including consumables and software, provides a simple, fast, end-to-end workflow for SMRT sequencing.
Although the Sequel system occupies a smaller footprint and is less than one-third the size and weight of its predecessor, it offers access to the key attributes associated with SMRT sequencing, including long reads, high consensus accuracy, uniform coverage, and integrated methylation information. Since the new system is built on the company’s established SMRT technology, most aspects of the sequencing workflow are unchanged.
“The system’s lower price and smaller footprint represent our continued commitment to leveraging the scalability of our technology and the unique characteristics of SMRT sequencing,” says Michael Hunkapiller, PhD, CEO of Pacific Biosciences. “Moreover, with its lower cost of goods (approximately a quarter of that of the PacBio RS II) we expect to be able to achieve substantial gross margin improvement and move more quickly toward profitability.”
“We will continue to support our PacBio RS II customers, and we expect to introduce improvements in sample prep, sequencing chemistry, and software that will extend the performance of that system,” says Hunkapiller. “We expect to make similar, substantial performance improvements each year for the Sequel system. In addition, the Sequel architecture provides the ability to scale throughput by substantially varying the number of ZMWs on future SMRT cells, thereby optimizing throughput and operating costs for specific applications.”
The Sequel system is designed for projects such as rapidly and cost-effectively generating high-quality, whole-genome de novo assemblies. It can characterize a wide variety of genomic variation types, including those in complex regions not accessible with short-read or synthetic long-range sequencing technologies, while simultaneously revealing epigenetic information. Using the company’s Iso-Seq protocol, the system can also be used to generate data for full-length transcriptomes and targeted transcripts.
“We are excited to support the human genetics community as they pursue the generation of higher quality whole human genomes, and move beyond SNPs to sequence the full size-spectrum of human genetic variation,” says Jonas Korlach, chief scientific officer at Pacific Biosciences. “With the introduction of our Sequel platform, SMRT sequencing will be available to more scientists seeking to find the underlying heritability of genetic diseases.”
According to the company, the Sequel system’s increased throughput should also facilitate applications of SMRT technology in metagenomics and targeted gene applications, for which interrogation of larger numbers of individual DNA molecules is important.
At the beginning of October, Pacific Biosciences’ SMRT technology was featured in 36 podium and poster presentations at the 2015 annual meeting of the American Society of Human Genetics, in Baltimore.
Currently designated for research use only, the Sequel system was developed as part of the company’s collaboration with F. Hoffman-La Roche Ltd, Basel, Switzerland, which is ultimately aimed at providing a nucleic acid sequencing system for use in human in vitro diagnostics. Under that agreement, Roche agreed to pay Pacific Biosciences $40 million in milestone payments related to the development of the Sequel system. The company previously reported that it has earned $20 million to date, and now expects to earn the remaining $20 million during the fourth quarter of 2015.
“This new sequencing platform has significant advantages over existing commercial platforms, and will be used as the basis for the Roche sequencing instrument being developed initially for clinical research, followed later by an IVD instrument launch,” says Dan Zabrowski, head of Roche sequencing and tissue diagnostics. “We anticipate the initial launch in the second half of 2016.”
Pacific Biosciences expects to begin limited US shipments of the Sequel system during the fourth quarter of 2015, and will begin scaling the manufacturing process for Sequel systems and new SMRT cells during early 2016. Shipments outside the United States are expected to commence thereafter. A portion of the initial group of Sequel instruments will be delivered to Roche to expand its internal assay development program.
HOW FAST? HOW SOON?
While commercial manufacturers of sequencing systems are proceeding at a measured pace from research use only to clinical application, academic researchers and innovators are continuing to make important advances that could dramatically improve future generations of sequencing systems.
At Washington University School of Medicine in St Louis (WUSM), a group of scientists have developed a new form of sequencing test that promises to detect virtually any virus that infects either people or animals. The metagenomic shotgun approach, called ViroCap, is designed to overcome the main difficulty that clinicians encounter when attempting to identify a disease-causing virus: the insensitivity of current methods to detect low levels of viral genetic material.
Currently, making the diagnosis of a viral illness can require clinicians to order a battery of tests to rule-in or rule-out specific infectious viruses. In some cases, current tests are capable of detecting only known viruses already suspected of being responsible for a patient’s illness. But with thousands of viruses as potential sources of disease, narrowing down the field can be difficult.
In September, study findings describing the ViroCap test were published online ahead of print in the journal Genome Research.2 Using patient samples, the study demonstrated that the ViroCap test can detect viruses not found by standard testing based on genome sequencing. The test could be used to detect outbreaks of deadly viruses such as ebola, Marburg, and severe acute respiratory syndrome (SARS), as well as such common viruses as rotavirus and norovirus, both of which cause severe gastrointestinal infections.
“With this test, you don’t have to know what you’re looking for,” explains senior author Gregory Storch, MD, professor of pediatrics at WUSM. “It casts a broad net and can efficiently detect viruses that are present at very low levels. We think the test will be especially useful in situations where a diagnosis remains elusive after standard testing, or in situations in which the cause of a disease outbreak is unknown.”
ViroCap sequences and detects viruses in patient samples and is just as sensitive as the gold-standard PCR assays that are widely used in clinical laboratories. However, even the most expansive PCR assays can only screen for up to about 20 similar viruses at the same time.
For the published study, the researchers evaluated their new method in two sets of biological samples (blood, stool, and nasal secretions) from patients at St Louis Children’s Hospital. In the first group, standard testing that relied on genome sequencing detected viruses in 10 of 14 patients. But the new test found viruses in four children that earlier testing had missed. Standard testing failed to detect common everyday viruses, including influenza B, a cause of seasonal flu; parechovirus, a mild gastrointestinal and respiratory virus; herpes virus 1, responsible for cold sores in the mouth; and varicella-zoster virus, which causes chickenpox.
In the second group, which encompassed children with unexplained fevers, standard testing detected 11 viruses in the eight children evaluated. But the new test found another seven viruses, including a respiratory virus called human adenovirus B type 3A, which is usually harmless but can cause severe infections in some patients.
Taking both study groups into account, the number of viruses detected by the new test jumped from 21 to 32—a 52% increase.
“The test is so sensitive that it also detects variant strains of viruses that are closely related genetically,” says corresponding author Todd Wylie, an instructor of pediatrics at WUSM. “Slight genetic variations among viruses often can’t be distinguished by currently available tests, and they complicate physicians’ ability to detect all variants with one test.”
In addition, because the test includes detailed genetic information about various strains of particular viruses, subtypes can be identified easily. For example, the study showed that while standard testing identified a virus as influenza A, which causes seasonal flu, the new test indicated that the virus was a particularly harsh subtype called H3N2.
Last flu season, H3N2 contributed to some 36,000 deaths in the United States. And in some patients—particularly young children, older adults, and people with weakened immune systems—knowing that the H3N2 strain is present may alter treatment.
To develop the ViroCap test, the researchers targeted unique stretches of DNA or RNA from every known group of viruses that infects humans and animals. In all, the research team included 2 million unique stretches of genetic material from viruses. These stretches of material are used as probes to pluck out viruses in patient samples that are a genetic match. The matched viral material is then analyzed using high-throughput genetic sequencing. As completely novel viruses are discovered, their genetic material can easily be added to the test, says Storch.
Coauthor Kristine Wylie, PhD, assistant professor of pediatrics at WUSM, investigates the viruses that set up residence in and on the human body, collectively known as the virome. The new test will provide a way to capture the full breadth and depth of such viruses, and deepen understanding of how they play a role in keeping the body healthy.
“It may also be possible to modify the test so that it could be used to detect pathogens other than viruses, including bacteria, fungi, and other microbes, as well as genes that would indicate the pathogen is resistant to treatment with antibiotics or other drugs,” she says.
The researchers plan to conduct additional research to validate the accuracy of the test, so it could be several years before it is clinically available. In the meantime, the technology can be used by scientists to study viruses in a research setting. The Washington University researchers are making the technology publicly available to scientists and clinicians worldwide, for the benefit of patients and research.
“In my view, the driver for the adoption of NGS in clinical applications is going to come down to cost reduction,” says Lewinski. “To make it possible for NGS to be more readily adopted in clinical settings, it will have to be demonstrated that the technology has the capability to perform diagnostic procedures more cost-effectively than alternative methods.”
In order to reap the benefits of improved patient outcomes or reduced hospitalization costs, laboratory directors sometimes find themselves needing to justify the purchase of an expensive technology that supplements but can’t replace existing equipment, says Lewinski. To do so, clinical utility and outcomes studies are essential.
“This is precisely the situation with NGS,” says Lewinski. “In order to benefit from the use of NGS systems in clinical settings, we’ll need to support adoption by conducting clinical utility and outcomes studies.
“NGS needs to gain its foothold in laboratories equipped to perform necessary quality control and standardization activities, and where there are experienced people to run the assays,” says Lewinski. “Correct implementation has to start with experienced users. The clinical utility and adoption of NGS will find their roots in those efforts.
“It’s certainly appropriate for academic centers and large reference laboratories to begin adopting NGS,” says Lewinski. “For those institutions, the time is now.”
Steve Halasey is chief editor of CLP. He can be reached via [email protected].
- Lefterova MI, Suarez CJ, Banaei N, Pinsky BA. Next-generation sequencing for infectious disease diagnosis and management, a report of the Association for Molecular Pathology. J Molec Diagn. 2015;17:1–12; doi: 10.1016/j.moldx.2015.07.004. Epub ahead of print, September 30, 2015.
- Wylie TN, Wylie KM, Herter BN, Storch GA. Enhanced virome sequencing using targeted sequence capture. Genome Research. 2015. Published online in advance, September 22, 2015; doi: 10.1101/gr.191049.115.