Nanopore Sequencing by Synthesis (Seq-by-Syn)

  • Yet Another Notable Achievement Involving George Church, ‘The Most Interesting Scientist in the World’ 
  • Team of 30 Coauthors Reports Seq-by-Syn with DNA Polymerase-Nanopore Protein Construct on an Integrated Chip
  • Challenging Improvements Needed for Commercial Reality

Prof. George M. Church. Taken from evolutionnews.org

Devotees of my blog will know that I’m prone to word play such as calling myself a “huge” fan of “tiny” nanopores for DNA sequencing, about which I’ve previously opined. They will also recall that I’m an admitted scientific admirer of George Church, who I think is The Most Interesting Scientist in the World.

Having said this, it’s not surprising that I closely follow what’s trending in nanopore sequencing, and also make an attempt to read all of Church’s papers as they get published because they are almost invariably quite interesting, involve “big ideas,” and in some new way are very educational, at least for me. Following are my comments about a recently published paper on nanopore sequencing in venerable Proceedings of the National Academy of Sciences of the United States of America (aka PNAS) wherein Church is the designated corresponding author.

Backstory

The seminal origins and early history of nanopore sequencing have been recently chronicled and criticized—then clarified—in Nature Biotech in several “To the Editor” items, which collectively provide enlightening insights into who did what when, so to speak. Those of us who are ‘Nanoporati’—a clever term tweeted by Nick Lowman—should definitely read those Nature Biotech items. For now, however, I’ll set the stage, as it were, by echoing a bit of what I’ve posted in the past for nanopores.

Patented but prophetic (i.e. no data) methods for nanopore sequencing DNA is actually a relatively old (~20 year) idea posited by Church and other creative visionaries. On the other hand, nanopore sequencing was first reduced to practice commercially not too long ago by Oxford Nanopore Technologies (ONT). Many years of delay between concept and commercialization was due to the need for gradual evolution of lots of “nanopore-ology” and sequencing biochemistry, as well as developing highly sophisticated electronics and complex algorithms for data analysis.

Nanopore Sequencing-by-Scanning (Seq-by-Scan)

Taken from rsc.org

As depicted below, and as can be best seen in a video, ONT’s commercially available MinION Seq-by-Scan system essentially involves threading a strand of DNA through a protein-based nanopore and converting resultant ionic current fluctuations into nucleotide base sequence.

While there are issues with base-calling accuracy, the remarkably small and readily portable MinION provides fast, real-time sequencing results for a wide variety of applications. These included unique or otherwise compelling Point-of-Care analyses, such as pathogen surveillance, which has been achieved in remote geographical locations and even in outer space aboard the International Space Station, as I’ve previously posted.

Nanopore Seq-by-Syn

In contrast to DNA Seq-by-Scan using a nanopore, which is challenged by pore-based differentiation of similarly sized A, G, C, and T bases, DNA Seq-by-Syn has no such limitation as it uses the DNA as a template for base-by-base (i.e. stepwise) detection of enzymatic synthesis of complementary DNA. Various Seq-by-Syn methods and challenges have been discussed elsewhere, and currently available commercial systems include those from Illumina and PacBio. The former employs nucleotides that are reversible terminators equipped with cleavable fluorescent “tags” on each base. The latter detects fluorescently labeled tags on polyphosphates released upon nucleotide incorporation.

The presently featured DNA Seq-by-Syn publication by Stranges et al., which builds upon two earlier reports cited therein, differs from the above approaches by using nanopore-based detection of mass tags rather than fluorescent tags. In principle, mass tags could afford higher accuracy compared to DNA Seq-by-Scan. However, as will now be explained, achieving improved accuracy is far easier said than done.

The general approach taken to demonstrate proof-of-concept for mass-tagged nanopore DNA Seq-by-Syn is depicted below in simplified cartoon form, but involves a true tour de force—in my opinion—of three key technologies. The first is design and synthesis of the nucleotides with appropriate mass tags, which involves very sophisticated chemistry that is best appreciated by reading detailed, extensive supporting information (SI) for Stranges et al. and SI for an earlier publication by Fuller et al. In a nutshell, these nucleotides have 5’-hexaphosphates linked to relatively large mass tags comprised of complex oligonucleotide structures.

Taken from Stranges et al. PNAS 2016

The second area of technical innovation involves attachment of a single molecule of ϕ29 DNA polymerase to each α-hemolysin (αHL) nanopore in such a manner as to retain its enzyme activity and be positioned such that every released mass tag transits through (i.e., is “captured” by) the nanopore leading to base identification by its current signature. As depicted below in two related representations, each of these heteroheptameric pores is comprised of one modified αHL subunit to which a peptidyl SpyTag moiety is attached, and six unmodified αHL subunits. This allows attachment of one ϕ29 DNA molecule modified with a cognate peptidyl SpyCatcher moiety at a predetermined, time-average distance from the pore.

Taken from Stranges et al. PNAS 2016.

The third key area of innovation deals with insertion of the enzyme-pore conjugate into a lipid bilayer residing on a silanized array (aka chip) of 256 Ag/AgCl electrodes such that there is one functional pore per electrode. Interested readers are encouraged to consult the publication for details, as well as check out related fabrication and methods patents that I found by searching Google Scholar.

Representative Results

The first image shown above depicts what base tag-specific detection would ideally look like if each of the four different bases would have a characteristic current-blockage intensity and persistence. In addition, all pores would ideally function similarly. Not surprisingly, given the stochastic nature of single-molecule systems in general, Stranges et al. found less than ideal behavior.

For example, out of 70 single pores obtained, 25 captured two or more tags, whereas only six of those pores showed detectable captures of all four tagged nucleotides. Data obtained for the pore with the most transitions between tag capture levels (i.e. the best results) is shown below, while results for the other five are given in the SI.

Taken from Stranges et al. PNAS 2016

To quote the authors:

“All four characteristic current levels for the tags and transitions between them can be readily distinguished…Homopolymer sequences in the template, and repeated, high-frequency tag capture events of the same nucleotide in the raw sequencing reads were considered a single base for sequence alignment. We recognized 12 clear sequence transitions in a 20-s period. Out of the 12 base transitions observed in the data, 85% match the template strand, showing that this method can produce results that closely align to the template sequence.” 

Interested readers need to consult and carefully read the SI for Stranges et al. regarding the interpretation of the “repeated, high frequency capture events,” such as that exhibited by C in the above current vs. time plot.

All of the above snippets in aggregate suggest to me that, while this huge amount of work has made progress toward one approach to Seq-by-Syn, many improvements will need to be made before achieving a robust system to successfully compete in the commercial sector.

Authorship, Affiliations, and Acknowledgments

The relatively large team of 30 coauthors listed for Stranges et al. include the following numbers of investigators and affiliations: 1 at Arizona State Univ., 4 at Harvard, 11 at Columbia University, and 14 at Genia Technologies, which is a Santa Clara, CA company that was acquired by Roche in 2014, and is part of Roche Sequencing.

Acknowledgments in Stranges et al. refer to support by Genia and NIH Grant R01 HG007415, which I found was awarded to coauthors George M. Church (Harvard), Jingyue Ju (Columbia), and James J. Russo (Columbia). The end of the abstract of this grant reads as follows:

“The nanopore chips will be enhanced and expanded from the current 260 nanopores to over 125,000 using advanced nanofabrication techniques. We will conduct real-time single molecule Nano-SBS on DNA templates with known sequences to test and optimize the overall system. These research and development efforts will lay the foundation for the production of a commercial single molecule electronic DNA sequencing platform, which will enable routine use of sequencing for medical diagnostics and personalized medicine.”

The conflict of interest statement in Stranges et al. indicates that the technology described therein (called “Nanopore SBS”) has been exclusively licensed by Genia, and that specified coauthors are entitled to royalties through this license. In addition, Church is a member of the Scientific Advisory Board of Genia.

Parting Comments

Long gone are the days when government-funded academic researchers thumbed their noses, if you will, at commercial development. Nowadays almost all academics parlay their government grants into university patents that get licensed to companies, usually with some type of corporate involvement of said academics.

I hasten to add that I’m not implying that NIH-funded academic research being a “seed” for corporate profitability is negative—especially in view of its Small Business Innovative Research (SBIR) program—but rather view it as a paradigm shift for the better, as it allows academic creativity to be harnessed into applications that can hopefully greatly benefit society.

In conclusion, and coming back to George Church, who I highlighted in the introduction to this blog, I must say that he might very well be the academic researcher with the longest list of technology transfer, advisory roles, and founded companies—13 to date—according to a public list that is truly mind boggling, at least to me.

As usual, your comments are welcomed.

Postscript

After writing this blog, Roche announced on December 15, 2016 that “it has officially notified Pacific Bioscience (PacBio) of its intention to terminate its [2013] agreement and efforts to develop a sequencing instrument for use in the clinical research and clinical market using their Single Molecule, Real-Time (SMRT®) technology,” about which I have commented previously. The announcement went on to say Roche would instead focus on internal development efforts” and “actively pursue multiple technologies and commercial strategies.” A GenomeWeb headline was more specific:  “Roche Will Focus on Genia’s Nanopore Technology for Dx Market After Ending Deal With PacBio.”

On December 30, 2016 it was reported that the University of California (UC) filed a patent suit against the Chief Technology Officer (CTO) at Genia, and Genia Technologies, claiming the CTO produced key inventions during his time at UC that he later assigned to Genia, but which should have automatically been assigned to UC. Stay tuned…

Frightening Fungus Among Us

  • Clinical Alert for Candida auris (C. auris) Issued by CDC
  • US Concerned About C. auris Misidentification and Drug Resistance
  • Sequencing C. auris DNA in Clinical Samples is Preferred for Identification
Strain of C. auris cultured in a petri dish at CDC. Credit Shawn Lockhart, CDC. Taken from foxnews.com

Strain of C. auris cultured in a petri dish at CDC. Credit Shawn Lockhart, CDC. Taken from foxnews.com

When I was a kid and didn’t know better, there was a supposedly funny rhyme that “there’s fungus among us.” While this saying is thankfully passé nowadays, the growing number of infections by a formerly obscure but deadly fungus is frightening. This so-called “superbug” is an antibiotic-resistant fungus called Candida auris (C. auris) that’s worth knowing about, and is the fungal focus of this blog.

First, Some Fungus Facts

Fungi are so distinct from plants and animals that they were allotted a biological ‘kingdom’ of their own in classification of life on earth, although that was only relatively recently, i.e. 1969. There are 99,000 know fungi, which exist in a wide diversity of sizes, shapes and complexity that extends from relatively simple unicellular microorganisms, such as yeasts and molds, to much more complex multicellular fungi, such as mushrooms and truffles.

It was previously thought that genomes of all fungi are derived from the genome of the model fungus Saccharomyces cerevisae, which has been used in winemaking, baking and brewing since ancient times. However, genome sequencing of more than 170 fungal species has revealed that, while the genome size of S. cerevisae is only ~12 Mb, seven species of fungus have genome sizes larger than 100 Mb. This is attributed to various evolutionary pressure-factors generating transposable elements, short sequence repeats, microsatellites, and genome duplication, and noncoding DNA.

Fungal cell walls are made up of intertwined fibers mostly comprised of long chains of chitosan, the same tough compound found in the exoskeletons of animals such as spiders, beetles and lobsters. The chitin in fungal cells is entangled with glucans and other wall components, such as proteins, forming a mass that protects the cell membrane behind it—and posing a formidable barrier against antifungal drugs.

Taken from Wikipedia.org

Taken from Wikipedia.org

In researching whether there are any nucleic acid drugs against fungi, I found one early patent by Isis (now Ionis) Pharmaceuticals for use of antisense phosphorothioate-modified oligonucleotides for the treatment of Candida infections, but virtually no other reports. I suspect that will change in the future as pathogenic fungi and other disease-causing microbes become more resistant to conventional drugs.

Fungal infections of the skin are very common and include athlete’s foot, jock itch, ringworm, and yeast infections. While these can usually be readily treated, infections caused by pathogenic fungi have reportedly risen drastically over the past few decades. Moreover, with the increase in the number of immunocompromised (burn, organ transplant, chemotherapy, HIV) patients, fungal infections have led to alarming mortality rates due to ever increasing phenomenon of multidrug resistance.

Segue to a Serious Situation

Emergence of drug-resistant fungi is, in part, the segue to the serious story of the present blog. The other part being incorrect identification of a certain fungus as being a common candida yeast, which is not only scary but seemingly inexcusable in today’s era of highly accurate PCR-based assays to accurately identify microorganisms. Here’s the situation in a nutshell.

  1. auris infection, which is associated with high mortality and is often resistant to multiple antifungal drugs, was first described in 2009 in Japan but has since been reported in countries throughout the world. Unlike many Candida infections, C auris is a hospital-acquired infection that is contracted from the environment or staff of a healthcare facility, and it can spread very quickly.

To determine whether C. auris is present in the United States and to prepare for the possibility of transmission, the Centers for Disease Control (CDC) and Prevention issued a clinical alert in June 2016 requesting that C. auris cases be reported.

(A) MALDI-TOF schematic; (B) mass spectra from three C. parapsilosis; and (C) two C. bracarensis isolates. Taken from researchgate

(A) MALDI-TOF schematic; (B) mass spectra from three C. parapsilosis; and (C) two C. bracarensis isolates. Taken from researchgate

This official alarm bell, if you will, was triggered by the following facts:

  • Many isolates are resistant to all three major classes of antifungal medications, a feature not found in other clinically relevant Candida
  • auris identification requires specialized methods such as a MALDI-TOF mass spectrometry or sequencing the 28s ribosomal DNA, as pictured below.
  • Using common methods, auris is often misidentified as other yeasts, which could lead to inappropriate treatments.

The CDC subsequently found that seven cases were identified in Illinois, Maryland, New York and New Jersey. Five of seven isolates were either misidentified initially as C. haemulonii or not identified beyond being Candida. Five of seven isolates were resistant to fluconazole; one of these isolates was resistant to amphotericin B, and another isolate was resistant to echinocandins. While no isolate was resistant to all three classes of antifungal medications, emergence of a new strain of C. auris that is would pose a serious public health issue.

Sequencing 28s ribosomal DNA. Taken from microbiologiaysalud.org

Sequencing 28s ribosomal DNA. Taken from microbiologiaysalud.org

Based on currently available information, the CDC concluded that these cases of C. auris were acquired in the U.S., and several findings suggest that transmission occurred:

  • First, whole-genome sequencing results demonstrate that isolates from patients admitted to the same hospital in New Jersey were nearly identical, as were isolates from patients admitted to the same Illinois hospital.
  • Second, patients were colonized with auris on their skin and other body sites weeks to months after their initial infection, which could present opportunities for contamination of the health care environment.
  • Third, auris was isolated from samples taken from multiple surfaces in one patient’s health care environment, which further suggests that spread within health care settings is possible.

A related Fox News story adds that C. auris was found on a patient’s mattress, bedside table, bed rail, chair, and windowsill. Yikes!

While the above situation in the U.S. might not seem particularly worrisome to you, the potential for emergence of more infectious C. auris strains with higher lethality should be of concern. That has already reportedly occurred in several Asian countries and South Africa. Obviously, deployment of the best available methods for pathogen identification can, in principle, lessen the likelihood of the emergence and/or spread of C. auris in the U.S. and other countries.

Case for Point-of-Care C. auris Nanopore Sequencing?

Taken from extremtech.com 

Taken from extremtech.com

Regular readers of my previous blogs know that I’m an enthusiastic fan of the Oxford Nanopore Technologies minION sequencer, which is proving to be quite useful for characterizing pathogens in very remote regions on Earth—and even on the International Space Station to diagnose astronaut infections! Notwithstanding various current limitations for minION sequencing of microbes, it seems to me that it would be relatively straightforward to generate minION data for many available samples of pathogenic fungi and genetically related microbes to assess the feasibility using minION for faster, cheaper, better unambiguous identification of C. auris minION in centralized or Point-of-Care applications.

Taken from rnaseq.com

Taken from rnaseq.com

If you think this suggestion is farfetched, think again, after checking out these 2016 publications using minION:

The 51.4-Mb genome sequence of Calonectria pseudonaviculata for fungal plant pathogen diagnosis was obtain using minION.

The first report of the ~54 Mb eukaryotic genome sequence of Rhizoctonia solani, an important pathogenic fungal species of maize, was derived using minION.

Sequence data is generated in ~3.5 hours, and bacteria, viruses and fungi present in the sample of marijuana are classified to subspecies and strain level in a quantitative manner, without prior knowledge of the sample composition.

CDC on C. auris Status and FAQs

In the interest of concluding this blog with the most up-to-date and authoritative information, I consulted the CDC website and found statements and replies to FAQs that are well worth reading at this link.

As a scientist, my overriding question concerns the lack of adoption of improved microbiological methods by hospitals and clinics. The above noted misidentifications of C. auris infections resulting from use of flawed lab analyses seems unacceptable. Although I don’t know all the facts or statistics to generalize, I suspect that there are other incorrect lab analyses due to use of outdated methods. On the other hand, I’m hopeful that, with the FDA’s widely touted Strategic Plan for Moving Regulatory Science into the 21st Century, the section entitled Ensure FDA Readiness to Evaluate Innovative Emerging Technologies—think nanopore sequencing—becomes actionable, sooner rather than later.

Changing established—dare I say entrenched—clinical lab tests is not simple or easy, but if it doesn’t begin it won’t happen, about which I’m quite certain. I can only wonder why development of infectious disease analytical methods and treatments seem to require a crisis. Sadly, I think it boils down to the complexities and socio-political dynamics of who pays.

Frankly, it’s my personal opinion that maybe it’s time Thomas Jefferson’s philosophy about hammering guns into plows is directed to health care.

Postscript

After writing this blog, I learned that T2 Biosystems has received FDA approval to market in the U.S. the first direct blood test for detection of five yeast pathogens that cause bloodstream infections: Candida albicans and/or Candida tropicalis, Candida parapsilosis, Candida glabrata and/or Candida krusei.

Yeast bloodstream infections are a type of fungal infection that can lead to severe complications and even death if not treated rapidly. Traditional methods of detecting yeast pathogens in the bloodstream can require up to six days, and even more time to identify the specific type of yeast present. The T2Candida Panel and T2Dx Instrument (T2Candida) can identify these five common yeast pathogens from a single blood specimen within 3-5 hours.

T2Candida incorporates technologies that break the yeast cells apart, releasing the DNA for PCR amplification for detection by greatly simplified, miniaturized nuclear magnetic resonance (NMR) technology, as can be seen in this video.

In my opinion, this fascinating new technology is another example of what could be rapidly deployed toward detecting C. auris.