Commentary
Homologous Recombination in H3N2 Seasonal Flu Confirmed
Recombinomics Commentary 06:55
June 1, 2008
HA sequences from approximately 13,000 H3N2 isolates were released today at Genbank. These sequences were generated under a broad consortium as listed for the Science paper, “The global circulation of seasonal influenza A (H3N2) viruses.”
Included in these sequences were examples of obvious recombination from patients in South Korea in 2002. These sequences had been generated by the Korean CDC and deposited at Genbank in 2004. Although the six sequences with recombination fell into two related groups, all sequences had been generated by the same lab, raising concerns of lab error.
The data had been presented on August 21, 2006 at the vaccine meeting, “Targeted Immunotherapeutics & Vaccine Summit” demonstrating (see slide 17) the recombination and showing the two groupings with A/Cheonnman/323/2002, A/Cheonnman/338/2002, A/Cheonnman/340/2002 in one group and A/Kyongbuk/320/2002, A/Daejin/258/2002, and A/Incheon/260/2002 in a second group. Although the two groups were easily distinguished, all six isolates had clear cut homologous recombination. The recombinants matched the contemporary sequence, A/Wyoming/03/2003 for the first 574 positions, and then switched to the 1991 isolate, A/Seoul/45/91, for positions 575-963. However, the first set switches back to contemporary sequences over short stretches.
A recent Journal of Virology paper, “Homologous recombination is very rare or absent in human influenza A viru,” found short stretches of recombination, but only found two examples of longer regions of recombination, which it attributed to lab error. The study severely limited the sequences analyzed, and excluded the six examples from Korea through a number of restrictions. The study only looked at sequences generated under an NIAID influenza sequencing program, which was limited to a small number of contributing institutions over a relatively short time frame for the vast majority of sequences. The study also required identification of the parental sequences, which had to come from the limited dataset.
The recombination in Korea was limited by a number of criteria. Only 2 of the 8 gene segments from the Korean patients had been sequenced, so in addition to not being sequenced under the NIAID program, the sequences were excluded because they where not generated by whole genome analysis. Moreover, the parental sequences were from human sequences circulating a decade earlier, and these sequences were also not in the NIAID program. Moreover, the public sequences were only of HA and were partial sequences. However, although the recombinant and parental sequences did not meet the criteria of the study, they did represent sequences that were either generated via homologous recombination or lab error.
The sequences in the Science paper confirmed that the data was not due to lab error. One isolate, A/Cheonnman/323/2002, representing the first group, was re-sequenced in Japan and the sequence matched the original sequence. Another isolate, A/Incheon/260/2002, was re-sequenced by the CDC in Atlanta and also matched the original sequence.
Thus, the original examples of homologous recombination in the six sequences generated by the Korean CDC, were independently confirmed by labs in Japan and the United States.
These confirmations demonstrate the need for broader analysis in the search for homologous recombination in human influenza. The first set of isolates also demonstrates multiple template switches, which decreases the size of the acquired sequences. These shorter sequences were found in the J Virology study, where the likelihood that the NA data was due to chance instead of homologous recombination was one billion to one. However, the most common exchanges happen between closely related sequences, resulting in acquisitions of single nucleotide polymorphisms, such as the example of G743A in NA of H5N1.
These acquisitions from previously described sequences allow for predictions of changes, leading to vaccine targets more representative of future emerging genomes.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings