By Kevin Davies
December 9, 2010 | In a dramatic piece of ultra-quick genetic detective work, next-generation sequencing company Pacific Biosciences has decoded the sequence of the strain of bacteria responsible for the deadly cholera outbreak in Haiti. The findings, which confirm the putative Asian origin for the devastating disease, are published online in the New England Journal of Medicine today.
The project was led by physician scientists at Harvard Medical School and Massachusetts General Hospital (MGH), including Matthew Waldor, John Mekalanos, Stephen Calderwood and Morton Swartz. “This understanding has important public health policy implications for preventing cholera outbreaks in the future,” says Mekalanos.
Cholera was first detected in Haiti in mid October, spreading across the country and into the Dominican Republic. Nearly 2,000 people have died from the outbreak, with no end in sight. Shortly after the outbreak, Waldor contacted the Centers for Disease Control and Prevention (CDC) and offered to sequence the bacterial strain using Illumina’s technology. Waldor says the CDC initially said he could have the strain, but five days later, changed their mind (citing political reasons) and said they were going to do it themselves. “At that point, I thought we were out of the game,” says Waldor.
CDC subsequently announced that using pulse-field gel electrophoresis fingerprinting technology, the strain was consistent with a south Asian origin. “But from a pure scientific point-of-view, that’s heresay,” Waldor, who is professor of medicine at Harvard Medical School and an investigator with the Howard Hughes Medical Institute, told Bio-IT World. “What are their controls? Pulse-field gel electrophoresis has nothing like the depth of a full genome sequence.”
But by then, Waldor and colleagues were already putting the finishing touches to their manuscript. Two weeks earlier, two MGH physicians -- pediatrician Jason Harris and Richelle Charles – returned from Haiti with samples they’d collected from a hospital. But who would do the sequencing?
That Was the Week
Two days earlier, on Saturday November 6, Waldor emailed a speculative inquiry to the PacBio website. “I knew they had some exciting technology, my understanding was it was very useful for resequencing bacterial genomes.” While he was fishing around on the PacBio Web site, Waldor noticed that one of his colleagues at the Brigham & Women’s Hospital – Joseph Bonventre – was on the PacBio advisory board.
“So I called him up,” Waldor continues. “He was in his office that Saturday, just like me, I told him the story, and he said, ‘let me make a phone call.’ Literally five minutes later, the CEO of PacBio, Hugh Martin, called me up, and said, ‘that sounds very interesting. Let me talk to Eric Schadt and my team.’ We got the strains on Monday November 8. Eric and the CTO called me that day and said they’d be interested in collaborating.”
“We’re going all in!” Waldor recalls Schadt telling him. “They went all in, I must say.”
Waldor’s team grew up the Vibrio cholerae strains on November 8, and the DNA samples arrived at PacBio in California on Wednesday, November 10. “We got a good idea of the [identity of the] two Haitian strains on the evening of November 12. We sent three other strains for comparison, including a true resequencing of the canonical strain.”
Each of the five strains took about one day to sequence to about 60X coverage. “They did an outstanding job in the analysis,” says Waldor. “Most of the credit for this project goes to Eric and his team.”
“The rapidity and depth of the sequence using this 3rd-generation sequencing technology has enormous potential to transform how we can analyze outbreaks of infectious disease and even the prediction of future outbreaks because of the power of their technology.”
According to PacBio, the five cholera genomes were sequenced on November 12 to 12-15X coverage in less than two hours. Further runs bumped up the coverage to 60X over the course of the day. Over the next three days, the sequence data were subjected to in-depth analysis, including genome assembly, annotation, and sequence comparisons, including comparisons to nearly two dozen published cholera genomes.
Subsequent bioinformatic analysis confirmed earlier hints matching the Haitian cholera outbreak to a variant of the “El Tor O1” variant from South Asia. This strain has never been documented in the Caribbean or Latin America, suggesting that a recent visitor to the island, possibly a volunteer or a United Nations peacekeeper helping relief efforts after the earthquake, could have inadvertently carried the bacteria to Haiti from outside Latin America.
“Our data strongly suggest that the Haitian epidemic began with the introduction into Haiti of a cholera strain from a distant geographic source by human activity,” is how Waldor puts it. The results disprove another possibility, namely that the strain arose from the local aquatic environment.
The identification of the Haitian strain has important implications regarding vaccination, says Waldor. “By showing this strain is closely related to a south Asian strain, and not close to Latin American isolates, it shows that human activities – food or water brought from South Asia, led to this epidemic and not from transfer from Latin America. That’s a conclusion that allows us to alter our policies in the future to prevent such a thing. For instance, relief workers or security forces should be deployed where there is no domestic or endemic cholera. Otherwise, workers should be screened and/or vaccinated, so they can’t bring it in.”
Speed of analysis is crucial in such situations, says Jason Harris, requiring a technology “that could immediately provide comprehensive genomic information about this virulent strain and quickly get it into the hands of the global health and research community. In the initial stages of a major epidemic, real time is the speed we need to be working in order to have the greatest impact on saving lives.”
From PacBio’s perspective, Schadt says that “real-time monitoring” of pathogens opens the door to using his firm’s technology as “a routine surveillance method, for public health protection in addition to pandemic prevention and response.”
Just last month, Waldor and colleagues published a perspective the New England Journal advocating the establishment of a cholera vaccine stockpile in the United States to be used to counter outbreaks such as the one in Haiti. There are an estimated 3-5 million cases of cholera each year resulting in about 100,000 deaths.
“The resistance to vaccination is truly baffling,” Waldor said at the time. The Harvard/PacBio results raise another troubling possibility: expansion of the epidemic with the replacement of the currently endemic strains with much more threatening variants. “That would be a deeply troubling outcome,” says Mekalanos. “A cholera vaccination campaign might not only control the disease but also minimize the dissemination beyond the shores of Hispaniola,”
The scientific manuscript was drafted over a few days and submitted to NEJM on November 19. The paper was formally accepted on December 1 and published December 9. “That’s like my record,” says Waldor.
Further Reading: Chin, C-S. et al. “The origin of the Haitian cholera outbreak strain.”New England Journal of Medicine December 9, 2010.