Sunday, 1 September 2013

Is there a better smoking bat or camel?

That teensy fragment of Middle East respiratory syndrome coronavirus (MERS-CoV) sequence (yes, I called it a fragment of that virus) from a Taphozous perforatus bat caused a lot of hassle last week,certainly a disproportionate amount to it's representation of only 0.5% of a MERS-CoV genome. Similarly, the report of MERS-CoV protein-reactive antibodies in camels some weeks back.

In a New York Times (NYT) article discussing the discovery of the 180-203 nucleotide (nt) gene fragment in a Saudi Arabian tomb bat, Donald McNeil opened with..



Health officials confirmed Wednesday that bats in Saudi Arabia were the source of the mysterious virus that has sickened 96 people in the Middle East, killing 47 of them.

The size range represents numbers used in various articles and of the fragment from the public sequence database GenBank (203nt) -"~190 nt" noted in the actual scientific publication). By the way, has so much ever before been written about so tiny a sequence?

Because he did not include in this line, or his article, a list of all the various possible scientific shortcomings, he didn't write in more detail about the difficulties with linking a virus in any sample to the cause of a disease  in humans, he forgot to specify that this was not the actual virion that caused MERS in the human index case in Bisha, he left out the PCR-101 section on why detecting a genetic sequence is not the same as isolating an infectious virus or how to interpret a PCR fragment's sequence....he was, in some circles, criticised. 

Okay, so the bat study did not isolate infectious virus, could not obtain any other sequence, found sequence in a bat from the family Emballonuridae rather than the "expected" Vespertilionidae bats and the positive sample was from bat faeces rather than blood or some other sample more convincing. Perhaps insects, food for T.perforatus, carry MERS-like CoVs? No-one has ever found that though. Is there evidence for 2 CoVs to be completely different except for a stretch of ~100% identity? Don't know, but don't think so. As Prof Andrew Rambaut noted in his very detailed analysis of this fragment, it does differ by 1 nucleotide from many MERS-CoV sequences (so its 99.5%-100% identical). Is there a more likely animal carrying a more similar MERS-CoV strain of virus that is spilling over to humans causing MERS cases? None that has been publicised to date. And therein likes my beef with some of the criticisms I have read this week. 

So far, this finding is the best lead we have in finding an animal source. There is no evidence to dispute the link, any more than there is evidence to prove it. Yes, there may be a another smoking bat or camel or something else out there. But it hasn't been found yet. And the public might like to hear how researchers are progressing rather than wait the very long time it will take to dot Is and cross Ts on the final MERS-CoV life-cycle, once they determine it.

And by the way, there is no other CoV sequence on GenBank that has >90% nucleotide identity with the T.perforatus sequence, except for the human MERS-CoV sequences. That doesn't exclude there being some other recombinant or novel CoV out there, but that is pure speculation; more so than saying that this new sequence represents a strain of the MERS-CoV found in humans.

In succinctly summarising some of the criticism, an article in CIDRAP presents a great overview and hints at what this criticism implies; that stories in the media must get every detail spot on or the writer may be portrayed as a poor scientist.

What? Wait. Seriously??

This was a newspaper article in the New York Times people. It was about 870 words long. It won't be setting global health policy nor will it be creating a WHO disease notification stating Taphozous perforatus bats, in particular, are the primary source of all MERS cases. Or of any cases. I don't doubt this is a prestigious newspaper but this story will likely be nest week's fish 'n chip wrappings (does anyone still use newspaper to wrap fish 'n chips? Is there a digital equivalent of - "yesterday's homepage, tomorrow's archive"?). In my opinion, and I don't mean to speak for all, scientists and health policy makers know that a newspaper is not a peer-reviewed scientific journal and that it's intent is to inform it's readers so they'll come back.

Were the many readers of the NYT misinformed? Perhaps the "health officials" could have been better defined by the NYT article. Presumably it's the authors-researchers may have been a better descriptive (as was used in a follow-up piece), probably more in tune with the public's perception of us. Beyond that the article did a good job of presenting the results of a research paper's relevant findings to the wider audience. They also both caution against over-interpreting the data. The NYT article has that well covered; a transmission route is not clear, more testing needed, more work being done, sample degradation due to a break in the cold chain, it was only 1 bat.

I very much agree with comments in Robert Roos CIDRAP article; the critics are getting carried away. There are many different levels of science communication that reach the general community - the popular press are not Lancet, and vice versa.

If you really want to pick holes in a part of the coverage, you might well ask why so many are looking at a 180-190nt fragment when the ends of that PCR-amplified fragment actually reflect the commercially made oligonucleotide "primers", not the (likely) viral template at all.  The actual fragment that should be analysed from the T.perforatus bat droppings is, at best, 156nt (but only 137nt if the internal nested PCR product was sequenced but the product on GenBank, 203nt long, includes both internal and external primer sequences in it-PCR speak here, sorry). Probably won't change the outcome of any analyses to date (156 still encompasses the nucleotide variation and is still differs by 11% from any non-MERS-CoV sequence), but it is a different number. 

I'm sure a newspaper headline "156 nucleotides of a 30,130 nucleotide genome possibly related to the mystery virus that may have directly or indirectly killed 47 people in Saudi Arabia" would be a real page turner.

So, let's keep up the good work of presenting and trying to deconvolute our own studies, let's keep the public interested and informed without overcooking the message, let's allow for imperfection (as readers of this blog will be all too familiar with), but perhaps let's keep the peer reviews to the scientific literature where they are in demand and required...and keep perspective on these new findings when they come to our attention through the popular press.