If you need more pointers, a useful starting point is Virus Res (2005) 107:141-149, which specifically explains why that error rate is a vast underestimate and gives a long list of determined error rates.

Oh, hell, I have it open anyway, I'll show it to you (I guess I can't include images in here, the table is here)

Please, do some real reading, and some thinking, before you post on this again. You are starting with a correct observation (the 40-ish per genome per year, on a population basis). Instead of using this to reject a fundamental fact of virology, you need to understand how the two observations are consistent, and what that means for natural selection on the virus. It's a very interesting and really profound connection, and you can't possibly understand influenza until you put these two facts together.

> All RNA viruses have roughly the same underlying mutation rate (~3 x 10^-5 errors per
> replication per base).

{are these nucleotides or amino acids ?}

3 times more than your 10^-4 but 10 times less than my 40/3/365/13000

You're simply wrong. This isn't guesswork, it's basic virology.

hmm, 1993. Now we have genbank.

Oh, come on. I cited a 1993 article to show this is basic and has been known for decades. This is absolutely basic, first-year-undergraduate, virology — RNA-dependent RNA polymerases are highly error prone and make, as a reasonable average, one error per genome per replication, and therefore RNA viruses exist as a quasispecies, a cloud, not a single sequence.

it can't be. That's simple logics and not virology and I'd bet on it.

It can be. It's simple math and simple virology.

Vincent Racaniello has a good primer on the basic virology you might find useful. Check out “The error-prone ways of RNA synthesis” and the posts before and after it. As he says there (and again, this isn't airy hypothetical stuff, it's basic, fundamental, undergraduate stuff):

Given a typical RNA viral genome of 10,000 bases, a mutation frequency of 1 in 10,000 corresponds to an average of 1 mutation in every replicated genome. If a single cell infected with poliovirus produces 10,000 new virus particles, this error rate means that in theory, about 10,000 new viral mutants have been produced. This enormous mutation rate explains why RNA viruses evolve so readily. For example, it is the driving force behind influenza viral antigenic drift.

No, you're mixing two very different things up here. We only see 40 (or whatever) mutations accumulate per year in the flu population, after selection at the population level and cycling through multiple hosts. But I'm talking raw mutation frequency, the error rate of the viral RNA polymerase; and that's very high. It's long been known that the error rate is so high, in fact, that essentially every replication incorporates at least one error. For example (one of many):

there is a clear central tendency for lytic RNA viruses (bacteriophage Qfi, poliomyelitis, vesicular stomatitis, and influenza A) to display rates of spontaneous mutation of 1 per genome per replication

–Rates of spontaneous mutation among RNA viruses. John W. Drake. Proc. Natl. Acad. Sci. USA Vol. 90, pp. 4171-4175, May 1993

That means that, far from 90% of viruses in a host being identical, essentially all flu viruses in a host are different.

Thanks for posting those links, but you'll notice that most of them are either things I've posted (including the post you're replying to), or papers I cited in those posts.

You're simply wrong. This isn't guesswork, it's basic virology.

hmm, 1993. Now we have genbank.

Oh, come on. I cited a 1993 article to show this is basic and has been known for decades. This is absolutely basic, first-year-undergraduate, virology — RNA-dependent RNA polymerases are highly error prone and make, as a reasonable average, one error per genome per replication, and therefore RNA viruses exist as a quasispecies, a cloud, not a single sequence.

it can't be. That's simple logics and not virology and I'd bet on it.

It can be. It's simple math and simple virology.

Vincent Racaniello has a good primer on the basic virology you might find useful. Check out “The error-prone ways of RNA synthesis” and the posts before and after it. As he says there (and again, this isn't airy hypothetical stuff, it's basic, fundamental, undergraduate stuff):

Given a typical RNA viral genome of 10,000 bases, a mutation frequency of 1 in 10,000 corresponds to an average of 1 mutation in every replicated genome. If a single cell infected with poliovirus produces 10,000 new virus particles, this error rate means that in theory, about 10,000 new viral mutants have been produced. This enormous mutation rate explains why RNA viruses evolve so readily. For example, it is the driving force behind influenza viral antigenic drift.

>> most flu-viruses (>90% in a host I guess) are still identical, genetically.
>> We only see ~40 mutations accumulated per year, one every 10 days.
>> Most mutations are synonymous and selection should have little effect.
>
> No, you're mixing two very different things up here. We only see 40 (or whatever)
> mutations accumulate per year in the flu population,
> after selection at the population level

which should be small for the majority of synonymous mutations

> and cycling through multiple hosts.

the virus won't know what host it is in

> But I'm talking raw mutation frequency, the error rate of the viral RNA polymerase;
> and that's very high.

very high = 3*10^-3 per position per year for flu-A, that's what you usually find.
I don't see why this can be much higher for a single replication cycle without
the mutations accumulating over the year

> It's long been known that the error rate is so high, in fact, that essentially every
> replication incorporates at least one error.

a replication in one cell may generate 10000 copies. How many of these contain at least
one mutation ? When you say : at least one, then I agree. But not more than 50%

> For example (one of many):
> there is a clear central tendency for lytic RNA viruses (bacteriophage Qfi, poliomyelitis,
> vesicular stomatitis, and influenza A) to display rates of spontaneous mutation of 1 per
> genome per replication
> –Rates of spontaneous mutation among RNA viruses. John W. Drake. Proc. Natl. Acad.
> Sci. USA Vol. 90, pp. 4171-4175, May 1993

hmm, 1993. Now we have genbank.

> That means that, far from 90% of viruses in a host being identical, essentially all flu viruses
> in a host are different.

it can't be. That's simple logics and not virology and I'd bet on it.

> Thanks for posting those links, but you'll notice that most of them are either things
> I've posted (including the post you're replying to), or papers I cited in those posts.

maybe. Just what I found when searching. Maybe I can complete/improve it later if you want.

IMO it's wise to be possibly wrong rather than to avoiding the subject. It should be discussed,
elaborated,clarified.

> I think that there's probably no such thing as two identical flu viruses
> (maybe I overstate, but not by much) – they exist as a quasispecies cloud.

most flu-viruses (>90% in a host I guess) are still identical, genetically. We only see ~40 mutations
accumulated per year, one every 10 days. Most mutations are synonymous and selection
should have little effect.

> So if flu infection starts with more than one virus, then you're probably being infected by
> multiple different viruses.

more than one = multiple. Usually 2, I guess.
10-1000 viruses may enter the body, I read, but most times they are all cleared. And most are
identical (>99%) since they found their way into the same droplet.
(I'm speculating)

> (However, note that in the case of HIV, which is certainly a quasispecies, it looks as if
> new infections usually do being with a single virus, and all the variability is lost at each
> new infection and has to start over).

seems to happen in flu also (>80%,IMO). Else we should see more mutations.
I don't know much about other viruses.

> But then recombination would look just like mutation — that is, the background mutation rate
> would be at least as likely to generate new sequences as would recombination of these
> two closely-related viruses.

call it reassortment for flu. It should create virus-triples with mutations at
(A but not B)
(B but not A)
(A and B)
which should be rare without reassortment. What's the freuency of these triples in different
virus databases ?

> So it's almost pointless to even think about the scenario of co-infection with
> closely-related viruses, because you're not going to get out of it anything that
> you wouldn't get out anyway. If that makes any sense.

the difference between mutations acquired by coinfection of different viruses in the same droplet
and mutation during replication should be the better balance of the concentrations.
Mutations that happen in later replication cycles have to compete with all the viruses
from previous replication cycles. One cycle = 6-10 hours.

some links that I found:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC274…
http://www.sciencedirect.com/science?_ob=Articl…
http://www.biomedcentral.com/1471-2180/3/11/
http://jcm.asm.org/cgi/content/abstract/48/2/369
http://jvi.asm.org/cgi/reprint/JVI.00773-09v1
a screening of the first 2000 influenza virus samples published on GenBank
for the IGSP (influenza genome sequencing project) show that approximately 3%
have some evidence of large-scale sequence polymorphism suggestive of
mixed infection.
http://www.iayork.com/MysteryRays/2009/08/28/in…
collection of over 1000 influenza samples. A plausible number they offer is about 0.5% of their samples — half a dozen individuals — were potentially mixed infections.5
http://www.iayork.com/MysteryRays/2009/05/18/on…
in fact, about 3% of the samples in the {flu-}database are contaminated
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC238…
Thus, up to 44 of 167 (26%) of isolates potentially represent mixed infections in the initial cloacal sample
Given the SLD procedure, the true rate of mixed infection, as defined by the presence of >1 HA and/or NA subtype, was likely to be

But then recombination would look just like mutation — that is, the background mutation rate would be at least as likely to generate new sequences as would recombination of these two closely-related viruses. So it's almost pointless to even think about the scenario of co-infection with closely-related viruses, because you're not going to get out of it anything that you wouldn't get out anyway. If that makes any sense.