How similar could be between Bacteria haplotype in term of SNPs?

Question

[jesselee516]

Hi all,

I had a question about the bacteria haplotypes(may be called strain? not sure here). We usually measure the SNPs for that haplotypes. For example, we have two haplotypes for one bacteria species, haplotype1 and haplotype2. How similarity could be these two haplotypes in term of shared SNPs ratio. In my research, I found two haplotypes may share 20%-40% of SNPs, can we say the shared number of two haplotypes are reasonable? Or at least 70%(some other number)?

[GenoMax]

This may be of interest: https://doi.org/10.1093/bioinformatics/btz280

Link for the software in paper is incorrect. Correct link: http://hulab.ucf.edu/research/projects/BHap/BHap.html

[Joe]

Haplotype terminology isnt really applied to bacteria since they’re for all intents and purposes haploid already.

I’m not sure I fully understand the question; are you asking how many SNPs one strain can have versus another to still be classified as the same strain/species?

[jesselee516]

Yes, 'are you asking how many SNPs one strain can have versus another to still be classified as the same strain/species?'. That is basically my question.

Or I can ask: 'normally, what percentages of SNPs are shared between strains under the same species'

Answer 1

[Joe]

Quite honestly, there’s no real answer to this, not least because the ‘species concept’ for bacteria is a clusterf* at the best of times.

Firstly, how many SNPs are sufficient depends hugely on the species/genus in question. Mycobacteria, for example, are (in)famous for being highly, highly genetically similar, in which case just a handful of SNPs would likely be considered sufficient (I’m not mycobacterium expert though so someone who is might have to correct me).

By comparison, in E. coli for instance, SNPs are commonplace, possibly the several 10s or hundreds at least. There is also the influence of mobile elements, which often contribute to strains as we know them (particularly if its a virulence plasmid).

Very, very approximately, a rule of thumb is that anything >95% average nucleotide identity is the ‘same thing’, but as I said, context is everything.

[jesselee516]

Thanks, your answer really make sense.