Genetic Comparison of Influenza A strains

[Paul_Bloom]

Hi all
I'm new here, but I'm currently working on a project for a computational biology class involving the evolution of the influenza a virus, and similarities between the genes of various pandemic strains.

Our hypothesis is currently that all strains of pandemic influenza A descend from the H1N1 virus that caused the Spanish Flu pandemic. Also, we feel that the closer a virus is (genetic distance) to the H1N1 virus, the more of a pandemic threat it may be. However, with more research I've considered that we may want to think that the H2N8 virus that caused the "Asiatic" pandemic of 1889 may be a better reference point than the H1N1 virus. I've also read some research that suggests the H2N2 virus is a recombination of the H2N8 virus.

Now, our methods are relatively simple, we use the program Matlab to get sequences of genes from NCBI, and locally align them, create phylogenetic trees from the Jukes-Cantor distance values, and then finally run a bootstrapping algorithm to assess the confidence of the tree branches.

We use two of the HA, NA, PB1 and PB2 genes for each viral strain. The current strains we use are H1N1, H3N2, and H5N1.

Our results are sparse, as that is pretty limited information. We decided to increase the number of genes to six per virus (6 HA, 6 NA, 6 PB1, 6 PB2). This is promising, but I have a few questions for Influenza experts.

What should be our 'baseline' influenza strain? Should we use H1N1, H2N8 (there's very limited data on NCBI available) or H2N2 (which is reportedly very similar or was the cause of the 1889 pandemic).

Should we add more influenza strains? I've considered throwing in the H7N7, H1N2 and H2N2, as well as the H2N8.

Just in terms of notation of the viruses, I've just assumed this, but if two strains share the same "H" or "N" notation, does that mean they have a certain mutation within the HA or NA gene? And if so, would it be valid to draw a conclusion that relates the relation of similar mutations (H1/2/3/etc) to the virulence of the strain, as well as its pandemic potential as compared to the 'baseline' strain?

I can add our inital phylogenetic trees as well as other specifics, but please help!

[HenryN]

Re: Genetic Comparison of Influenza A strains

Hi all
I'm new here, but I'm currently working on a project for a computational biology class involving the evolution of the influenza a virus, and similarities between the genes of various pandemic strains.

Our hypothesis is currently that all strains of pandemic influenza A descend from the H1N1 virus that caused the Spanish Flu pandemic. Also, we feel that the closer a virus is (genetic distance) to the H1N1 virus, the more of a pandemic threat it may be. However, with more research I've considered that we may want to think that the H2N8 virus that caused the "Asiatic" pandemic of 1889 may be a better reference point than the H1N1 virus. I've also read some research that suggests the H2N2 virus is a recombination of the H2N8 virus.

Now, our methods are relatively simple, we use the program Matlab to get sequences of genes from NCBI, and locally align them, create phylogenetic trees from the Jukes-Cantor distance values, and then finally run a bootstrapping algorithm to assess the confidence of the tree branches.

We use two of the HA, NA, PB1 and PB2 genes for each viral strain. The current strains we use are H1N1, H3N2, and H5N1.

Our results are sparse, as that is pretty limited information. We decided to increase the number of genes to six per virus (6 HA, 6 NA, 6 PB1, 6 PB2). This is promising, but I have a few questions for Influenza experts.

What should be our 'baseline' influenza strain? Should we use H1N1, H2N8 (there's very limited data on NCBI available) or H2N2 (which is reportedly very similar or was the cause of the 1889 pandemic).

Should we add more influenza strains? I've considered throwing in the H7N7, H1N2 and H2N2, as well as the H2N8.

Just in terms of notation of the viruses, I've just assumed this, but if two strains share the same "H" or "N" notation, does that mean they have a certain mutation within the HA or NA gene? And if so, would it be valid to draw a conclusion that relates the relation of similar mutations (H1/2/3/etc) to the virulence of the strain, as well as its pandemic potential as compared to the 'baseline' strain?

I can add our inital phylogenetic trees as well as other specifics, but please help!

The H and N numbers are "serotypes" and based on reference antisera, but all H1's will be closer to other H1's than any other H, etc. However, the definition of a serotype would involve a large number of shared positions.

H1N1 caused the 1918 pandemic, H2N2 was 1957, and H3N2 was 1968.

[AlaskaDenise]

Re: Genetic Comparison of Influenza A strains

Why not include H9N2, since it has also infected humans?

Does your study involved all species, or are you looking only at humans? If humans only, why?

How do you define pandemic? - only novel antigens, or novel antigens with high path mutations (NS1/92 + polybasic cleavage site + ???)
Origins of novel antigens are easier to explain.

Also - welcome to the forum!

.

[gsgs]

Re: Genetic Comparison of Influenza A strains

1889 was assumed to have been H3N8
not all panflu descended from 1918 H1N1 in all segments.
1957 caught avian HA,NA,PB1 1968 caught avian HA,PB1

no recombination with H2N2 or H2N8 AFAIK

see flugenome.org for classes of flu-viruses for each segment

I can post a list of mutual differences in the segments of some important ancient viruses

[sharon sanders]

Re: Genetic Comparison of Influenza A strains

Welcome Paul!

[Paul_Bloom]

Re: Genetic Comparison of Influenza A strains

Thank you all for the help! I was reading wikipedia about the 1889 pandemic, so I can see they have something wrong when they say H2N8. I will instead use the H3N8 virus.

Really, my group and I wanted to use those three viruses because we knew we could find dozens of each gene sequence on NCBI. I know there are lots of H2N2 sequences as well, but we're still working on finding enough of the other sequences I mentioned.

I'd really like to be able to continue this as research, with recombination - I work in a computational biology lab, but we use Perl/Python to compare genomes of animals mostly, not so much with viruses. Hopefully the other molecular/cell bio courses I take (including virology) will let me do some sort of research like this.

Thanks for the warm welcome! Just a few more questions
What are the consequences of recombination of the previous pandemic flu viruses? Is it possible that they will 'fizzle' and be susceptible to current antivirals? I read over a paper posted on here a while ago that looked at recombination events between varied influenza A viruses and said that about 50% of them were 'more effective' at replicating within a host, and several strains were much more virulent.

See: http://www.flutrackers.com/forum/showthread.php?t=69339 for the paper.

[gsgs]

Re: Genetic Comparison of Influenza A strains

I'm using C, only flu - no other organisms. I'm looking for people to exchange
(processed/improved) data with.
No recombination involved in pandemics,
many experts doubt there is recombination at all in flu.

No mention of recombination in that PLOS-paper, only reassortment
(exchange of whole segments)

[Laidback Al]

Re: Genetic Comparison of Influenza A strains

Welcome, Paul. You have come to the right place to pursue your research interests.

As you can see, there are several members here that are more than willing to answer your questions. They are also more than willing to share their opinions as well.

[Paul_Bloom]

Re: Genetic Comparison of Influenza A strains

Ok -
To expand our research, we expanded our data to H5N1, H3N2, H1N1, H2N8, H3N8, H7N7 as main groups. We have the influenza B sequence to compare to, as well as other viruses that share HA, NA, PB1 and PB2 genes.

We locally align the sequences of the HA, NA, PB1 and PB2 genes, and have generated trees from that alignment.

I wanted your opinions on whether we should perform Ka/Ks analysis to help increase or data, or maybe global alignment as well. I can post images of our trees later today when I'm on the computer they're saved to.

Any opinions on whether we chose good genes to prove our hypothesis that the Spanish Flu (H1N1) is the most common ancestor to the other flu strains we've used?

[gsgs]

Re: Genetic Comparison of Influenza A strains

what's ka/ks ?

1918 H1N1 is not an anchestor of today's avian flu

[Paul_Bloom]

Re: Genetic Comparison of Influenza A strains

Ka is the rate of nonsynonymous mutations in a comparison of two genes, Ks is the rate of synonymous mutations. When you run this, it'll give you a graph that shows on a certain scale on the y axis, of zero to some number (3, 5, 10), that is the rate of Ka or Ks, and along the x axis it shows you the region of the sequence you're comparing, from the first nucleotide to the last nucleotide.

The Ka/Ks ratio indicates regions that are under positive selective pressure, that mutate rapidly, and typically any region with Ka/Ks > 1 is under such pressure.



That image shows the Ka/Ks along the sliding nucleotide window, and as it increases, either Ka increases or Ks decreases, meaning that you have fewer synonymous mutations or more nonsynonymous mutations...

[HenryN]

Re: Genetic Comparison of Influenza A strains

Ok -
To expand our research, we expanded our data to H5N1, H3N2, H1N1, H2N8, H3N8, H7N7 as main groups. We have the influenza B sequence to compare to, as well as other viruses that share HA, NA, PB1 and PB2 genes.

We locally align the sequences of the HA, NA, PB1 and PB2 genes, and have generated trees from that alignment.

I wanted your opinions on whether we should perform Ka/Ks analysis to help increase or data, or maybe global alignment as well. I can post images of our trees later today when I'm on the computer they're saved to.

Any opinions on whether we chose good genes to prove our hypothesis that the Spanish Flu (H1N1) is the most common ancestor to the other flu strains we've used?

1918 H1N1 is fairly unique. It is a recombinant between human H1N1 and swine H1N1. You can take Iowa swine from 1930/31 and human WSN/33 and find almost all polymorphisms in all 8 gene segments of 1918 H1N1,

[Paul_Bloom]

Re: Genetic Comparison of Influenza A strains

didn't gsgs say that recombination didn't occur in flu? i read a lot of work that supports what you said, and i think the amount of recombination in H1N1 supports what my group has found so far...

[HenryN]

Re: Genetic Comparison of Influenza A strains

didn't gsgs say that recombination didn't occur in flu? i read a lot of work that supports what you said, and i think the amount of recombination in H1N1 supports what my group has found so far...

Influenza has a segmented genome and reassortment is common. However, recombination is even more common, although most recombiantion is between closely related sequences, so the polymorphisms acquired by recombination look like point mutations.

Current influenza dogma maintains that antigenic drift is due to random mutations, while shift is due to reassortment.

Recombination is ignored or discounted as much as possible, but the growing sequence database is reaching critical mass (examples of recombination held on hard drives notwithstanding).