RapTOR seeks to quantify unknown unknowns
By Patti Koning
The world of bioterrorism is filled with scary stuff — anthrax, smallpox, and ricin, to name a few pathogens. And those are just the agents we know about. There is also a whole other realm of “unknown unknowns,” lethal agents that could be weaponized from ordinary viruses or disguised to look harmless.
The RapTOR (Rapid Threat Organism Recognition) Grand Challenge seeks to solve the “unknown unknowns” problem by developing a tool to rapidly characterize a biological organism with no pre-existing knowledge.
“We’ve been thinking about this threat space for years,” says Todd Lane (8623), the project’s principal investigator. “Taking advantage of rapidly evolving molecular biology technology and the advent of ultra-high-throughput DNA sequencing, we are re-engineering time-intensive benchtop methods to be faster, easier, and automated.”
Todd divides the biological threat spectrum into three categories: traditional agents such as anthrax; enhanced agents that have been genetically manipulated for increased virulence, drug resistance, or to evade detection; and advanced agents — the unknown unknowns. “An advanced agent could start with a benign organism that we’d have no interest in from a national security perspective and manipulate it into a virulent pathogen that is difficult to detect with current systems,” he says.
Unlike known threats such as anthrax or smallpox, detection of an advanced agent would only occur when people begin showing symptoms. Every day that treatment is delayed, the lethality of the attack goes up exponentially. “If a novel attack occurs and our detection systems fail, we have limited time in which to identify and characterize the organism to be able to offer effective treatment,” says Todd.
History shows that identifying and characterizing a naturally occurring unknown organism is very difficult. The 1970s outbreak of Legionnaires’ disease took six months to characterize; nearly 30 years later, it still took weeks to characterize Severe Acute Respiratory Syndrome (SARS). Conventional DNA sequence-based detection systems failed to identify a recent outbreak of Ebola in Uganda because the virus had changed so much it was unrecognizable.
Lowering the bar to bioweapons
The same advances that make the RapTOR concept feasible also have lowered the technical bar for creating a bioweapon. “The research I did in graduate school for my dissertation is now being taught in high school,” says Todd. “It is now possible to completely synthesize bacterial genomes. Bioweapons have become a potentially low-cost weapon of mass destruction and it’s a very risky situation.”
Sequencing the human genome took 10 years and hundreds of millions of dollars. With ultra-high-throughput sequencing, that same work can be accomplished for about $10,000 in a week or less.
But ultra-high-throughout sequencing only addresses part of the problem. As Todd explains, in an outbreak scenario there would be a large number of samples from people manifesting symptoms of the disease and the worried well.
“The more samples you can sequence, the better chance you have at identifying and then characterizing the organism. But sequencing a clinical sample provides a lot of information that is not of interest,” he says.
For example, 99 percent of the DNA in a blood sample is the human genome. DNA in a nasal swab is 90 percent human-derived and much of the rest is garden-variety bacteria. “You need to quickly eliminate the ‘human flora’ before sending a sample to ultra-high-throughout sequencing,” says Todd. “We aren’t exactly looking for a needle in a haystack — we’re looking for multiple needles and each one is different.”
A better analogy is a jumble of 100 different disassembled pocket watches, with one of the original pocket watches representing a pathogen. “You have to sort through the watch parts to identify and discard everything you recognize. You have to simplify the mixture so there are more parts of significance to the pathogen,” he explains. “It’s not enough to just identify those parts that are unique — you also have to reassemble the pathogen watch.”
Molecular biologists employ a number of methods to prepare samples for ultra-high-throughput sequencing; the challenge for the RapTOR project is adapting those methods to a portable, automated platform. Todd explains that these methods require days of work by a highly trained scientist on the bench.
“Our overall goal is a 24-hour turnaround. An end user would inject a sample of blood into the system, which then runs the sample through a number of molecular biology manipulations and sends it off to a DNA sequencer,” says Todd. “The sequencer provides enriched information that the user sorts through to identify and characterize the sample.”
The team already has succeeded in adapting one method to a microfluidic platform: normalization, which removes high-abundance genetic material from a sample, leaving a small, representative amount of all of the genetic material found in a sample. Currently, normalization is performed using an enzymatic digestion process that relies on enzymes from Siberian Kamchatka crabs.
A faster, cheaper, and simpler method
To re-engineer normalization, the researchers looked back 20 years to hydroxyapatite chromatography, a resin-based method that was discarded because it was difficult to reproduce. “Modern resins are commercially available and well-defined, so hydroxyapatite chromatography is relevant again,” says Todd. “We created a capillary-based system to perform hydroxyapatite chromatography. It’s a faster, cheaper, and simpler method that doesn’t destroy the material in the process.”
In the traditional, benchtop normalization process, double-helix DNA is heated to separate the two strands. As the DNA cools, the genes that are expressed in high abundance will find their partner strands more quickly than those expressed in low abundance. A researcher stops the cooling sequence, adds the crab enzyme to remove the double-stranded DNA, and with some additional manipulation, the resulting DNA all appears in low abundance.
Hydroxyapatite chromatography follows the same process, substituting a phosphate buffer for the crab enzyme to remove the double-stranded DNA without destroying any of it. The entire process can be automated.
The RapTOR normalization method is now being tested against human clinical samples for “fevers of unknown origin” that did not develop beyond mild sickness. “These are outbreaks that get ignored because they are self-limiting, but the samples are a perfect test of our system,” says Todd. “If we can handle a small outbreak, we can handle something larger.”
The RapTOR normalization method has caught the attention of sequencing companies and government agencies that Todd says are very interested in helping bring such a device to market. He expects a prototype to be ready in the fall.
RapTOR must do more than just normalize samples, so the researchers are turning their attention to other DNA manipulation methods such as ligation (linking two small pieces), digestion (cutting a larger piece into fragments), and size-based separations. Adapting some of these methods will prove simpler than others. Sandia’s microfluidic platform was developed for proteins, which are far more complex than DNA samples.
The LDRD project brings together a broad distribution of technical disciplines, unusual for a Grand Challenge but well-suited for Sandia’s capabilities. “Everyone has to work together, the microfluidicists, the microbiologists, and the bioinformaticists,” says Todd. “As we process samples, the data analysis and knowledge discovery group will provide feedback on the quality of the information being produced. The upstream process is tunable.”
RapTOR is designed to be a public health tool applicable for day-to-day work. “If you develop a tool like this, you need to regularly apply it to real-world scenarios. You can’t put it in a glass box and wait for the horrible attack,” says Todd.
The tool will have other applications in other fields as well, such as environmental detection. Todd explains that RapTOR could be used to take regular atmospheric samples and analyze their genetic makeup to develop a baseline. This application has garnered interest from the Defense Threat Reduction Agency, DoD, and DHS.
RapTOR will never eliminate the problem of unknown unknowns, but it will make the path from unknown to known faster and simpler, says Todd.
RapTOR for algae: Understanding pond collapse
Last month, Todd Lane (8623), Jeri Timlin (8622), and Ben Wu (8125) received $800,000 in funding over two years from the DOE Biomass Program for their proposal “Pond Crash Forensics.” Using pathogen detection and characterization technologies developed under the RapTOR Grand Challenge, they will compare the environmental conditions and metagenomes of algal samples taken from normal ponds to those taken from ponds that have undergone collapse.
Algae are widely viewed as a potential source of renewable fuel, but the technology to mass-produce fuel-grade algae is still in the early stages. A major roadblock, says Todd, is the inability to produce large amounts of algae.
Algae are commonly grown in raceway ponds, large, shallow, artificial ponds that serve as fields for algae crops. “Pathogens and viruses fall into these ponds and can crash a pond overnight,” says Todd. “No one has identified many of the agents that are causing these pond crashes. You can’t develop countermeasures without understanding why something is happening. This is a complex problem with a lot of factors at play.”
He adds that this is a mostly unexplored area because growing algae is closer to farming than biotechnology. “This is a good application for RapTOR because, like clinical blood samples, there is a lot of naturally occurring stuff to sort through before you can find the pathogen or virus,” says Todd. “It’s a really good niche for Sandia, to provide a service that will be of great benefit to the algal biofuel industry that will in turn greatly benefit the nation.”