The Demise of Random Mutations
Recombinomics Commentary 21:11
March 28, 2008
The upcoming paper, with the unlikely title, “Homologous Recombination is Very Rare or Absent in Human Influenza A Virus” signals the beginning of the end of one of the basic tenets of influenza genetics, Genetic Drift via selection of random mutations. This basic tenet has guided research and vaccine development in influenza genetics and is an underlying principle of evolution in general.
For influenza and other single stranded RNA viruses, evolution is thought to be driven by a series of random mutations generated during the copying of the virus’ genetic information. The errors which provide a selection advantage would then be incorporated into the viral genome, and these changes would lead to a drift of genetic information. These changes would allow the virus to escape from immunological attack, and lead to frequent changes in vaccine targets.
Since changes by random mutation are thought to be unpredictable, new vaccine targets are selected after the virus has evolved. Of course this leads to vaccines that are already somewhat out of date by the time they are available for use.
The concept that the changes were due to recycling of prior selected mutations through recombination represented a paradigm shift that was thoroughly discounted by influenza geneticists, including the authors of the above paper. They had previously argued that changes in the 1918 pandemic virus were due to differential evolution of the HA gene, and that homologous recombination in negative sense RNA viruses was rare and played little or no role in viral evolution in general and influenza evolution in particular.
The above paper represents the results from an analysis of human influenza sequences. The search was in a limited database of complete human sequences, which eliminated the detection of recombination in partial sequences, or recombination that involved donor sequences influenza isolates from birds or swine. Thus, the number of examples would be significantly lower than the true instance. Although only two examples of recombination involving long stretches of recombined sequences were found, there were over 300 examples of small stretches of homolgous recombination which were statistically significant. The significance for the 240 examples in the NA gene had a p value of <1.2 X10 to the minus 10. Thus, the likelihood that the result was not due to chance was more than a billion to one.
Since the evidence for small regions of recombination was strong, the authors have focused on lab error to explain the examples. They cite amplification of contamination or mixed sequences to create lab generated recombination. However, they provide no evidence that any of the more than 300 short examples are due to such contamination. Moreover, examples of recombination in shorter human influenza sequences, which were excluded from this paper, as well as examples in swine and avian influenza, contain examples of recombination which are present in multiple isolates in multiple locations.
The statistically significant examples are the beginning of the end of the use of random mutations to explain influenza evolution. As additional sequences are generated, the random mutation explanation for rapid evolution becomes less tenable. For sequences from H5N1 genes, the number has risen markedly in recent years. Most newly acquired polymorphisms are readily detected in the database, and most are in other H5N1 sequences. Moreover, as the number of sequences rise, more shuffling of individual polymorphisms is found, especially when individual polymorphisms are traced. These tracings through the sequence database generate clear patterns, which can be used to predict new combinations. Moreover, the same polymorphisms can jump from one genetic background to another.
Recombination is frequently between closely related sequences, which result in the stable transfer in a limited number of changes per recombination event. These changes are typically mistaken for single nucleotide polymorphisms linked to copy errors. However, the statistically significant patterns described in this paper will lead to more extensive analysis of these polymorphisms, which will lead to the understanding of the true nature of these genetic changes, which are driven almost exclusively by homologous recombination.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings