Post Publication Hoarding of H5N1 Sequences by The WHO
Recombinomics Commentary
November 21, 2006
Most journals now expect that DNA and amino acid sequences that appear in articles will be submitted to a sequence database before publication.
The above instructions to submitters of sequences to Genbank reflect the requirements for most major peer reviewed journals. The underlying premise of peer review publications require authors to include information needed for independent confirmation of the results being published.
In the past, WHO and consultants have withheld H5N1 sequences until publication. This hoarding of data prevents independent analysis of a rapidly evolving H5N1 genome. Such hoarding is curious, since the hoarders frequently receive samples for independent confirmation of H5N1 per WHO requirements. Thus, although the samples are collected by public health agencies, the data is withheld from the public by the WHO affiliated labs until publication.
However, the sequence hoarders have now extended the hoarding beyond publication. This has been done in two recent high profile publications.
One publication, in the Proceedings of the National Academy of Science (PNAS) on the spread of the Fujian strain of H5N1 in southern China, has sparked considerable controversy and requests by the WHO for China to release more H5N1 samples and sequences.
The second publication, in Nature, focused on changes in the receptor binding domain that increased affinity for human receptors. Both papers are heavily dependent on sequence data, and both papers use data that is being hoarded the the WHO private database.
The private database has thousands of withheld H5N1 sequences. One of the largest hoarders has been Weybridge, which received H5N1 samples from birds, mammals, and humans collected throughout Europe and the Middle East during last season's H5N1 outbreak.
Sequences from one bird in Turkey, four patients in Turkey, and one patient in Azerbaijan have been released in the past few months. However, a presentation of the spread of H5N1 in Europe in 2006 included phylogenetic trees (here and here), which contain approximately 80 HA Qinghai sequences that had been completed by the presentation date, May 30, 2006. Since there are 8 gene segments per isolate, and the presentation did not included human isolates, it is likely that well over 1000 sequences are currently being hoarded by Weybridge, even though many samples were collected over a year ago, and others were from the beginning of 2006.
The sequences in the PNAS paper were from 406 H5N1 isolates collected in southern China in 2005 and 2006. 406 isolates would generate 3248 gene sequences, but only 556 were released (404 full or partial HA sequences and 152 PB2 sequences). The paper discusses the genotypes of the isolates, which require sequence data for all eight gene segments from each isolate. The paper also discusses amantadine resistance, which is based on the sequence of the M2 gene. The paper includes accession numbers for the 556 sequences described above. There were no sequences released for six gene segments (PB1, PA, NP, NA, MP, NS).
The Nature paper described receptor binding domain changes in various H5N1 isolates from Vietnam and Thailand patients. Although the paper included the amino acid changes of the southeast Asian clones, the nucleic acids sequences from these clones were withheld. The paper repeatedly states that these changes were "mutations", but fails to provide sequence data that would include the silent changes in the RNA, which would contain data relevant to the "mutation" claim.
One of the changes discussed in the paper was S227N, which was found in one of the clones from a patient from Vietnam. This change had been predicted based on recombination, and the prediction on the precise nucleotide change was made prior to the detection of such changes in Turkey and Egypt. These isolates had the predicted nucleotide change, which was confirmed by the nucleotide sequences deposited by others at GenBank. The precise change could not be determined from the protein sequence, because of the redundancy in the genetic code. The change in the Vietnamese clone, or an earlier isolate, A/VN/JP12-2/05, from another Vietnam patient, can not be determined, because these sequences are being hoarded by WHO in its private database, even though papers on both sequences have now been published (the earlier isolate was published by the WHO Global Influenza Surveillance Network in Emerging Infectious Diseases).
In addition, the Nature paper cited detection of N186K and Q196R in Qinghai H5N1 isolates from patients in Azerbaijan and Iraq. These changes could not be confirmed, because sequences from these patients are also being hoarded in the WHO database.
Thus, the WHO is currently hoarding more H5N1 sequences than any agency, yet it requests samples and sequences from China, because of the importance of such sequences in the monitoring of the evolution of H5N1, which is required for the creation of updated diagnostic primers, as well as new vaccine targets.
WHO consultants have expressed the view that such vaccine targets could not predicted because the changes in H5N1 are due to random mutations. However, the public H5N1 sequences, as well as other influenza serotypes, have clear examples of recombination, which create new gene sequences by using previously identified changes. Similarly, the single nucleotide changes, which are considered as point mutations by the WHO consultants, can also be readily identified in reported sequences, including H5N1 sequences that are likely to be involved in dual infections via transport and transmission by wild birds.
As the sequence data accumulations, the "random mutation" explanations becomes less tenable. The NIAID has a flu sequencing project which will generate full sequences at no cost, yet the H5N1 sequences database is littered with partial sequences from published H5N1 collected since 1999. The generation and release of full sequences from these isolates would provide a more complete evolution of H5N1 and provide more examples of obvious recombination.
The hoarding of the sequences by WHO and consultants continue to be hazardous to the world's health.
The post-publication hoarding has set a new transparency low.
Media sources
Phylogenetic Trees