The Indonesian government has given permission to release the sequences of some 40 or so H5N1 viruses isolated from human cases in that country to publicly accessible gene databases. The US CDC has now removed any password protection preventing general access at the Influenza Sequence Database at Los Alamos National Laboratory and has indicated the sequences will also be deposited in the US National Library of Medicine's database, GenBank.
The move to put the data in the public domain, giving scientists from around the world free access, came after the Indonesian government told the World Health Organization on Thursday that it was willing to share the genetic sequences of all H5N1 viruses isolated from humans there.
"We feel this is a public health problem of global proportion and that therefore it is in the best interest of global public health to have the data available for as many creative minds and well-trained minds as possible," Dr. Nancy Cox, director of the CDC's influenza division, said in an interview with The Canadian Press. (Helen Branswell, Canadian Press)
In an exclusive interview, Cox told Branswell she hoped the Indonesian decision would induce other governments to follow their example:
She didn't name names. But the government of Turkey has never agreed to share the genetic codes of viruses from an outbreak in January. And China has been criticized in the past for sharing only a few of its isolates, and slowly at that.
"It's a very positive step on the part of the Indonesian government and ministry of health. And I hope that other countries will follow suit," Cox said.
The sequencing of the Indonesian isolates was been done at CDC and The University of Hong Kong laboratory of Malik Peiris. It is unclear whether Peiris has data not released by CDC, and if he does when he will release Indonesian sequences in his possession not otherwise public. We urge him to do so immediately, if he has not already done so. Every day from this point onward Peiris does not release such sequences will reflect badly on his currently excellent reputation as a scientist. It also remains unclear whether CDC has sequence information of its own, not subject to constraints by outside parties, which it has yet to release. If it does, keeping these sequences to themselves would be a gross act of hypocrisy.
And then there is the major issue of the password protected WHO database at Los Alamos which contains an unknown number of sequences not available to the general scientific community. This database is not only an affront to the ideal of open science but is especially offensive in the context of a possible emergence of a pandemic strain of influenza virus. Bioinformatic techniques are accessible to scientists in almost any country. They do not require high tech laboratories or specialized personnel. Just a standard computer and internet connection. This means a large number of good scientists can attack this problem from many different perspectives. We believe a major advance is just as likely, or even more likely, to be achieved by some "outside the box" scientific approach as by the extremely accomplished but more conventional scientists who now control the data like arrogant mafiosi. If the problem is lack of permission by particular governments, then they should say so and name names.
This is a new and dangerous world and the old rules don't apply.
- Log in to post comments
Revere: Now that the Indonesian sequences are public domain, when do you think we should realistically expect interpretive analysis to be released by scientists?
Nancy: Hard to say. The purpose in our minds for making them public is to allow a vastly larger group of scientists to work on them in conjunction with other sequences. Science is still slow. We are like the virus. We try this. Then we try that. It takes time. Could be tomorrow. Could be 6 months. Could be never.
I'm tough to please so it is not often that I say "good for them" to a government. But that being said; good for them to the Indonesian government.
are there any updates with regards to the etiology?
"Could be tomorrow. Could be 6 months. Could be never."
Just like the possibility of an H5N1 pandemic. I'm betting somewhere in the range of 6 months to never.
So now i can make my very own H5N1 virus in my basement lab? Cool.
My basement lab currently has a sophisticated device that can turn dirty clothes into clean. Not up to the standards of the philospher's stone, but pretty cool by itself, and it gets alot of use.
Stephen: You can't make it but you can analyze the sequences however you wish. Let us all know what you find out.
Again Revere you do not know what you are talking about. The sequences in the hand of HK have been released onto the LANL site, just as for the CDC, go and look. Of course more sequences will probably be forthcoming from both labs as they receive and sequence additional isolates sent to them by the Indonesian government. So the appearance of new sequences does not indicate hoarding.
You name someone as endangering public health without any facts or basis.
Shame Revere.
In the What's New section of his website, Niman has already posted s number of commentaries which can clearly be regarded as containing interpretive analysis.
Among other things, in an August 7 post he says that there are a large number of polymorphisms present in the human sequences that are not present in Indonesian avian sequences. In the same paragraph, I think he goes on (there is his usual attention to spelling and proofreading) to say that the presence of these polymorphisms are additional evidence that the majority of human H5N1 cases in Indonesia are not linked to avian H5N1 in Indonesia.
Is there anyone of sufficient scientific background to either agree or disagree with these two statements by Niman?
As yet, there are no sequences from avian isolates from Indonesia since mod-2005. Therefore it would be difficult to assess any genetic changes as there is no background with which to make a comparison. Apparrently the Indonesian government has sent avian isolates to Australia for sequencing, so hopefully this includes recent isolates from the same regions that human infection has been occurring.
On the naming of sequences. Time for a change?
This seems as good a place as any to interject a comment on nomenclature. The general structure of names e.g. A/Indonesia/5/05 H5N1 and its inconsistent application present problems when searching for data.
The structure Type/species/region/#/year (serotype) seems logical enough in theory but if you wish to search any of the fields the only one you can rely on is Type (A, B or C).
Species is, by convention, omitted for homo sapiens and may otherwise be a Latin name, common name or other (e.g. water sample, environment etc).
Region is often a country but may be a province, state, town etc.
# is normally just a sequential number but sometimes includes other data e.g. the name of the processing lab.
Year varies in format and may be yyyy or yy
Serotype normally sticks to the HxNx format but for HA sequence may just have Hx etc and may or may not be bracketed.
This may seem a petty gripe but if you which to generate a phylogenic tree for Indonesian sequences searching on Indonesia will not pull up sequences if a region has been used (looking on the bright side it has been good for my world geography - do you know where Mahachkala, Munia, Deli Serdang & Anhui are?). It is difficult to be sure you are looking at all the available data based on almost any database query. With the rate of change in flu viruses it is also limiting having data resolved only to the year, if you search for Bar headed goose & Qinghai in a given year are you looking at birds before they set off for the west or returning birds with new polymorphisms? Some country are enormous covering many time zones so a Lat/Long or GPS fix might be better as any identifier of sample collection.
A suggestion Type/species/Lat-Long/country code/#/mmyyyy (HxNx)
Some variations on the theme
A/chicken/Navapur/Maharashtra/India/7966/2006
A/chicken/Thailand/Kanchanaburi/CK-160/2005
A/cameron/46
With human H5N1 in mind some way matching the Accession number to clinical data e.g. did the patient recover, drug resistant?, part of a cluster? Strong evidence of poultry infection etc would also be helpful.
To make matters even worse I think munia is a type of songbird, and I believe a relatively commonly kept caged one also.
SSAL: I agree 100% with Henry. What we don't know is the other source of the recombinants.
The release of sequences has had a dramatic shift in the past week. Much of the new data is being discussed at flutrackers.com (I tries linking the discussion early, but teh link flags the post and the last attempt never was "approved").
Last Friday Indonesian sequences from both the CDC and Hong Kong were released. That was followed by three sets of releases yesterday. Hong Kong (Malik Peiris and Yi Guan) provided data on the more recent isolates (through collections on July 6, 2006) and added the second collection from the nephew in the Karo cluster.
Hong Kong also released the HA and NA data on the H5N1 from wild birds and poultry in Hong Kong this year. Those sequences matched the Fujian strain (in birds and people in China in 2005 and 2006 as well as birds in Laos and Malaysia in 2006). Hong Kong earlier released the Laos and Malaysia data before either country admitted there was an H5N1 problem - Malaysia still has not acknowledge teh seqeunce data.
Australia has also released data on 2005 poultry isolates from Indonesia. They weree sent 91 recent samples because the human and poultry sequences don't match. They have released HA and NA sequences, which clearly show that the poultry still does not match. A few were closer to the CDC bird sequences, which were from October 2005 (and also don't match). No 2006 bird sequences have been released yet, but it looks like the releases are almost coming in real time.
These releases are coming via Los Alamos because that is where the WHO private database is located. The earlier "releases" was simply the removal of the password protection - the sequences are generally deposted within days after receipt, which is within days of the date of death. The earliest deposit was July 29, 2005 (over a year ago).
As far as Turkey goes, four human sequences were released weeks ago, but requests for the age and gender of the patients involved have gone unanswered. Weybridge should have more than the 4 released sequences (they were sent 21 positives abd could confirm 12). They also have on the order of 100 sets of H5N1 sequences from birds in Europe. So far they have released one (the first isolate in Turkey from the fall of 2005).
So far no word out of Thailand, Laos, Vietnam on H5N1 in birds or people (and Malaysia has yet to acknowledge the Fujian sequences from the beginning of this year).
Although Indonesia and WHO are pushing hard for a domestic poultry link to the human sequences, it is unlikely that the effort will succeed. The first sequences sent to Australia have come out, and they have no match. The CDC also came out with October, 2005 chicken sequences, and they also did not match.
The newly released human sequences are loaded with mammalian and Qinghai wild bird polymorphisms. There are NO wild bird or mammalian sequences from Indonesia, other than the one cat sequence (which IS a match).
Contact with dead or dying poultry is the criteria being applied for H5N1 testing in Indonesia and southeast Asia (as in Thailand).
WHO's focus on domestic poultry is hazardous to the world's health. The Indonesian sequences are out and they clearly point AWAY from domestic poultry.
The poultry are the "canary in the coal mine", but NOT the source of many or most of the H5N1 infections in humans.