I am looking at the genome of the influenza virus being used in the flu vaccine 2017-2018: A/Michigan/45/2015.

When I look below at the individual segments, there are 4 sequences for each segment - shouldn't there be only one for a single "strain"? Is this due to rapid mutation rates? Or is it that they are including both positive and negative sense of the virus?

Edit: sorry, link here:



2 Answers 2


It looks to me that the website you're looking at is mostly taking publicly available data from Genebank (the NCBI repository for sequence data) an collects it to form an overview.

For your question this means that there are 4 available sequence datasets for the A/Michigan/45/2015 influenza strain - they are likely very similar but based on the reported segment length not quite identical.

Or is it that they are including both positive and negative sense of the virus?

If you mean the positive and negative strand of the DNA, then no - since the sequence of these two is always complementary the Genebank repositories only stores the forward sequence.

You can also see that the 4 different datasets have similar Acession numbers:
KY117020 - 27 (Passage M1/C3)
KU509700 - 07 (Passage E1)
KU933490 - 07 (Passage E3)
KY090607 - 14 (Passage M1/IR)

You can look up more details for these accessions on the genebank website:

  • It seems that all 4 datasets came from the same publication
  • ... which apparently was never published (so it's not available online), but it might be part of some WHO program?
  • The main difference of the 4 datasets I can find is the passage of the strain. I'm no virologist, so I'm not quite sure what it means, but in cell biology passage describes the time a certain cell line has been in (in vitro) culture.

This is pretty typical for influenza, and reflects the ambiguity of "sequence" when talking about rapidly-mutating RNA viruses. Such viruses exist as a "cloud" of variable sequences; a single replication of influenza is likely to contain variants at every single site of the genome. That also allows the virus to undergo extremely rapid natural selection, including during laboratory passages (which are obviously not performed using infected humans, but might be in various cell types, or very often in eggs).

A common scenario for an influenza isolate (especially one that, like MI/45, is used in vaccines) would be to swab someone's nose, use the swab to inoculate an egg to prepare a stock virus, send that stock to several other labs, grow a subset in cells, have another lab grow some in their cells or eggs, and so on. If it's used in vaccines, it might be made into a reassortant with a different (less virulent) backbone, but the surface genes would still be termed MI/45. Each lab would sequence the virus, often at multiple steps in the passage; and at each step, there's an opportunity for the virus to either undergo selection, or simply to drift, and change its sequence. Or, of course, if may not change, and identical sequences might be deposited into the databases, perhaps starting at slightly different points depending on each lab's sequencing protocol.

Many viruses are only sequenced once, perhaps as the original isolate or as an early passage, because there's no reason (or capacity) to look further at the tens of thousands of isolates that are taken every year, but where there's a reason to look at an influenza virus more than once, standard protocol is to re-sequence it each time it's used. I can guarantee that there are far more than just 4 sequences of MI/45 out there, probably there are more like 400 that have been done world-wide, it's just that there's no point in depositing all of them.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .