The “laboratory error” hypothesis, a review of the Gibbs paper

When the new H1N1 pandemic virus was first described in April, its unusual constitution was immediately noted. The 8 different genomic segments did not match any known virus. Instead, different genomic segments matched flu viruses that infected birds, pigs and humans. As internationally known flu scientist Robert Webster put it:

“Where the hell it got all these genes from we don’t know,” says Robert Webster, a flu virologist at St Jude Children’s Research Hospital in Memphis, Tennessee. “But this is a real super-mixed-up virus.”

Nature News, April 27, 2009

Ever since then, people have been looking for “the smoking pig” – a pig infected with a virus similar to the one causing pandemic flu. Thus far, this has been a futile search.

Failing finding a good match with a known virus, some researchers have analysed the sequences we do have to try to reconstruct a possible pedigree for the new H1N1 virus. The problem with these reconstructions is that they must account for the fact that the genomic segments best match viruses on different continents. This problem is usually explained as due to transportation of infected pigs from one continent to another.

There is another possibility. Pandemic H1N1 could have originated in a laboratory. It is routine to artificially combine influenza viruses in ways that could produce a hybrid virus similar to the new H1N1 virus. When I first read about the strange background for the pandemic virus, from bird, pig and human viruses, I immediately thought that it might have come from a laboratory. I was not the only one.

Adrian J. Gibbs, a well-known Australian virologist, first suggested that pandemic H1N1 may have come from a laboratory in May (Bloomberg, May 13, 2009). Predictably, the World Health Organisation (WHO) immediately rejected any possibility of this hypothesis. At the time, it was unclear what data prompted Dr. Gibbs’ hypothesis. A paper describing his analysis was published yesterday (November 24, 2009). What follows is a summary of some of the key points made by Dr. Gibbs and his collaborators.

Problems with the natural model

The assumptions and statistical methods applied by proponents of the natural origins hypothesis are discussed. These are highly technical. For our purposes, it is sufficient to point out that the confidence that one can assign to phylogenetic trees is proportional to the amount of sequence connecting key points in the evolution of the organism. In the case of pandemic H1N1, the natural hypothesis assumes that many years worth of “connecting” sequence data is missing. Hence, important corroborating sequence data for the natural hypothesis has not been found.

The natural model also assumes that pigs from Eurasia and North America shared viruses which produced a triple-reassortant virus, the pandemic H1N1. However, the “avian-like” viruses that circulate in Eurasia and are found in the new H1N1 have never been reported North America until the pandemic virus was identified. The North American parental viruses have never been seen in Eurasia. Hence, there is no empirical support for the idea that these two parental virus sources have had an opportunity to mix in nature.

The pandemic virus is suggested to have arisen in Mexico. The first person to have been reported with the disease lived in Perote, a small Mexican village. Much has been made of the fact that there are pig farms in the immediate vicinity. However, Perote is in an isolated valley, 200 kilometers from Mexico City, and surrounded by mountains. The authors point out that this remote location is an unlikely ground zero for the mixing of North American pigs with Eurasian pigs fresh off an intercontinental jet. Further, some of the other early cases in Mexico were in Oaxaca, 290 kilometers to the south of Perote. Finally, since phylogenetic analsyses suggest that the virus first infected humans in January 2009, the fact that the virus was first reported in a child in Perote, Mexico in April is of uncertain significance.

Evidence in favor of the laboratory error model

Occam’s razor, entia non sunt multiplicanda praeter necessitatem, suggests that the simplest explanation is the most likely to be true. In the case of pandemic H1N1, the simplest explanation for its odd parentage is that the marriage was consummated in a laboratory. Combining different flu viruses artificially is frequently done when studying the biological properties of different genes and when creating vaccines. The Gibbs paper focuses primarily on the idea the pandemic H1N1 virus might have originated from a vaccine, although he does discuss other possibilities related to influnenza research.

There are a number of sequence anomalies related to the pandemic H1N1 virus. In particular, the authors focused on the PB1-F2 gene. From the paper:

It seems that the peculiarities of the S-OIV PB1-F2 gene, the human-like signature sites and its selectively super-imposed termination codons, probably reflect the outcome of selection…

I will discuss sequence anomalies associated with pandemic H1N1 identified by Gibbs et al as well as ones I have observed in upcoming blogs.

Evidence against the vaccine hypothesis

The authors looked for evidence that the virus had sequence changes in the NS1 protein that would be consistent with preparation for an attenuated vaccine. Specifically, attenuated live vaccines have NS1 proteins truncated to 126 amino acids. Pandemic H1N1 has an NS1 protein that is 219 amino acids long. Thus, one potential line of evidence clearly does not support the hypothesis that the new H1N1 virus was developed as a live, attenuated vaccine.


Due to the controversial nature of this paper and the different ways that it could be interpreted, it is important to be clear on how the authors themselves interpret the results. First, they do not state that they have definitive proof that the pandemic H1N1 virus was born in a laboratory. They list problems associated with the natural hypothesis, provide evidence consistent with a laboratory origin and ask for more intense surveillance so that any missing sequences can be found. Second, Gibbs and his colleagues also do not use the word “bioweapon” at any point in the paper. Although they discuss several possibilities, they appear to be leaning towards some sort of error that occurred during the manufacture of flu vaccine for swine.

I am glad that Dr. Gibbs has been given the opportunity to publish his analysis and conclusions. I do not necessarily agree with all of his interpretations, but considering the full range of possible explanations for the pandemic H1N1 virus is the first step in finding the truth.


Gibbs et al. (2009) From where did the 2009 ‘swine-origin” influenza A virus (H1N1) emerge? Virology Journal. 6: 207


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s