Commentary
New Biology of Recombination in Influenza Confirmed
Recombinomics Commentary 15:50
June 26, 2008
we would predict that in most cases involving apparent “recombinants”, a new sequencing assay would result in a different sequence, since it seems unlikely that the reverse transcriptase jumps would occur in the same positions as before.
The above comment is from an accepted paper, “Anomalies in influenza genome database: New biology or lab errors?” which will be published in the Journal of Virology. It has much in common with an earlier paper in the same journal, entitled “Homologous recombination is very rare or absent in human influenza A virus”. The earlier paper focused on human influenza, but required identification of the parental sequences as well as the recombinant, and used a limited database, so the vast majority of the examples were of short sequences. However, the number of examples exceeded 300 and the p value was very low (for NA the value was less than 1 in a billion). However, although the number of examples was high, the authors speculated that these examples were due to lab error and the sequences were generated during amplification.
The new paper examines a wider range of samples, which include swine and avian sequences and it also identifies are large number of examples of obvious homologous recombination, but this paper also suggests that the sequences are due to lab error. They suggest re-sequencing of these recombinant sequences, because a new sequence would not match the initial sequence because the polymerase would nit jump from one template to another in the same points, so although a recombinant sequence would be generated by re-sequencing, but the sequences would not match because they would have variable contributions from the sequence in the sample and the sequence from the contaminant.
However, the database already had examples from six patients in Korea which were obvious recombinants and fell into two sets of three similar sequences. Thus three of the sequences began with the contemporary 2002 H3N2 HA sequence in circulation, and then switch to a sequence in circulation a decade earlier. However since the same sequence was from three patients, the recombinant was re-sequenced in one sense because all three samples yielded the same result. This was repeated in the second series of three samples, which jumped from the contemporary sequence at the same position, but then jumped back and forth between the two distinct sets of sequences. Once again, the similarity between all three sequences ruled out a contaminant.
The absence of a contaminant was recently confirmed by recently released sequences from two independent labs, which were published in Science. One sequence, A/Incheon/260/2002, was generated by the CDC in Atlanta, four years after the sequences were deposited by the Korean lab, and the sequence generated in Atlanta was an exact match. Similarly, a lab in Japan also took one of the Korean samples and also generated a sequence, A/Cheonnman/323/2002, that exactly matched the sequence generated four years earlier in Korea.
Thus, the original set of six sequences, as well as the re-sequencing by two independent labs four years later, confirmed that the sequences were not generated by template switching during amplification, but represented template switching by the two viruses in circulation in Korea in 2002.
Re-sequencing has also been done for H5N1 sequences from China, which also had obvious examples of recombination. Like the human sequences, the re-sequencing of the H5N1 produced recombinant sequences which were virtually identical to the initial sequences.
Thus, the re-sequencing of these isolates confirms the “new biology”, which is homologous recombination in influenza.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings