Hello, and thank you for listening to the MicroBinfeed podcast. Here we will be
discussing topics in microbial bioinformatics. We hope that we can give you some
insights, tips, and tricks along the way. There's so much information we all
know from working in the field, but nobody writes it down. There is no manual,
and it's assumed you'll pick it up. We hope to fill in a few of these gaps. My
co-hosts are Dr. Nabil Ali Khan and Dr. Andrew Page. I am Dr. Lee Katz. Andrew
and Nabil work in the Quadram Institute in Norwich, UK, where they work on
microbes in food and the impact on human health. I work at Centers for Disease
Control and Prevention and am an adjunct member at the University of Georgia in
the US. Welcome to the Microbial Bioinformatics podcast. Lee and Nabil are your
co-hosts today. We're navigating the twisted world of bacterial taxonomy. We
have some excellent guides to help us. Our guests today are Dr. Leydon
Pritchard, who is a Strathclyde Chancellor's Fellow at Strathclyde Institute of
Pharmacy and Biomedical Sciences at the University of Strathclyde, and Dr. Conor
Meehan. Dr. Conor Meehan is an assistant professor in molecular microbiology at
the University of Bradford. He specializes in whole genome sequencing and
molecular epidemiology of pathogens, primarily mycobacterium tuberculosis and
genome-based bacterial taxonomy. Welcome. Hi. Thanks for having us. So, every
time we have new guests on the show, we ask an easy question like, what do you
normally do for your work? And so, Leydon, what does it take for you? Other than
administration, which is pretty normal in academia, I do quite a wide range of
computational biology. For the last 10, 15 years or so, it's mostly been
microbial genomics. Prior to being at Strathclyde, I was at the James Hutton
Institute just outside Dundee, where I worked on plant pathogens and human
pathogens, which were associated with plants, so E. coli and salad greens, that
kind of thing. So there's been a lot of assembly, annotation, transcriptomics
and diagnostics and that kind of thing. And that's really the diagnostic sort of
things for legislations, where the interest in taxonomy came from and the need
to get involved in taxonomy. And it's why Pioneer, which is possibly five people
at the time I was working in this area, that's why that was written. But I'm
also interested in molecular level and systems level interactions, so I was
interested in how pathogens interacted with plants, and that now at Strathclyde,
because I've moved more towards biomedical and industrial biotechnology, that's
moved towards more medical-relevant pathogens in terms of environmental
bacteria, and that's moved towards streptomyces. And I've even started working a
little bit on human biology, looking at cardiomyocyte response to stress with
metabolic modelling. Excellent. And Connor, we haven't had you on for a while,
anything new with you? I guess I haven't talked much about the taxonomy side and
where my background and interest is at the moment. So my first interest in
taxonomy was I needed a project during my postdoc, and we had these
Lachnospiraceae that were found in the human gut. And me and my boss, Rob Biko,
at the time were like, this is interesting, let's look at these. So in a non-
hypothesis-driven way, we did a data-driven way and then found that there was
environmental niche differences in it. And the main taxonomy issue that was
there was that things had been renamed so often that they were very difficult to
find what was supposed to be in this genus and what wasn't. And now a lot of my
work is on the mycobacteria. So coming from a clinical taxonomy point of view,
what name should we use so that conditions can very quickly assign some kind of
treatment to it? When should we change that name so that it is more biologically
correct? When should we leave it so that it doesn't confuse people? So it's a
very applied taxonomy that I kind of come from. Let's start off with a difficult
one. So I wanted to ask both of you, I'm very confused when it comes to
taxonomy. I understand a species for higher organisms, but what about a species
for bacteria? I would start off by saying you probably don't understand how to
define a species in a multicellular either. The first time I ever came across
taxonomy was a lecturer, Grace McCormick, in my university in Maynooth where I
was studying my undergrad. We were first years and she split us into a group and
she said, come up with a species concept that covers everything, including
bacteria and viruses. And we thought this would be easy. And now here I am 20
years later or whatever, and we still don't have one. In multicellular, it tends
to be about reproduction and whether offspring are then viable for further
reproduction. And that's a very generally sweeping statement that does not
involve any plants. I think across prokaryotes and multicellular organisms, the
sort of abstract principle I fall back to is that there's some kind of
recognisable, if not barrier to gene flow, at least some kind of discontinuity
where it's not just a case where two organisms can exchange genetic material
easily and have that progress into the next generation. And that's going to be
defined differently for prokaryotes than it is for multicellular organisms just
because of what they are and how they interact with each other to reproduce. As
my background comes from looking primarily at genome sequences, I'm really only
looking at the genetic material. I'm looking at the flow of that
computationally. I'm not really looking personally very often at the larger
biological picture. So I always picture it in terms of some sort of disruption
or discontinuity in the ability to exchange genetic material. So if you want to
define a new species, though, and you actually want to do what people think most
taxonomists do, which probably is not even true, it is still primarily
biochemical. There will be a suite of biochemical tests, and then you are saying
this is the profile for this. It can grow on this. It does not grow on that. It
has this kind of colony. It has these kind of morphologies, this kind of
motility, all of these things. That still really is a basis for defining a
species. We're moving slowly away from that. But you need to have that strain.
It must be culturable. It must be deposited in two different collections. So
there are the ATCC, and there are the other ones that are around, or the
specialized ones for different species or organism groups. And then genetically,
at the moment, it's still based on 16S. So to get a little bit into the 16S side
of it, 16S has been around for 30-something, 40-something years now. And it was
found by Carl Woese and others that if you use the sequence of this gene and you
compared it, you would approximately get species groups that we already knew. So
you have to remember that a lot of taxonomy is we defined these groups before,
and then we use this marker to pull back those groups. So it was not done
agnostically to create new groups. And it was saying that 97% cutoff of the
whole sequence similarity, if it's within 97% of each other, this was
approximately equivalent to a 70% DNA-DNA hybridization in the lab, which was
approximately equivalent to the biochemically and physically defined species. So
everything still came back to that original biochemical, physical way. But then
we see these kind of groupings. And now what a lot of modern taxonomy is, is
that true? What does that mean? What do we do with that information? And I think
the historical aspect is really important, because what you described there, I
think, as we knew that some organisms could be divided into species, and then we
kind of almost retrofitted 16S back onto that to aid with a sort of quantitative
classification, is kind of what we've done with genetics onto organisms, which
we recognize as being distinct in historical terms. We knew that, say, cats and
dogs were distinct from species, but we didn't really understand the genetics
until relatively recently. And we've always kind of followed this historical
trend, whether it's this categorization that we've made as humans and labeling
of things in the environment and around, or whether it's taking the biochemical
and phenotypic taxonomy of microorganisms and then trying to retrofit our
genomic analyses onto that to try and make them match up with what we considered
to be species previously. So taking that sort of broader philosophical view, I
think there's always been a sort of historical tendency to try and remain within
what went before. And that's true for, say, binomial nomenclature as well, where
the binomial nomenclature is a particularly good way of describing the way that
bacteria mix their DNA and divide and pool off into distinct reproducing groups.
But it's how the Linnaeans evolved within culture, and we try to retrofit back
to that. I would definitely attend your lectures if I could. I intend to go off
on a philosophical side, because I used to work for Ford Doolittle, and that's
what we would talk a lot about, is the philosophy of trees and how that tree
then creates a nomenclature that is binomial, that then from that creates
lineages. And what does it mean to be a lineage? Yeah, that was my Thursday
afternoons. Tell me a funny Ford Doolittle story. I only know about him by name.
I will tell you a funny one, which is when I worked at Dalhousie University in
Canada, I was funded on a CIHR grant, which was to look at microbiomes in 2010.
And actually, we had a work package on defining what is the reproductive unit of
a prokaryote. And there is every answer, and there is no answer to that. But
that was an actual work package that we had. We had philosophers that were built
on that, and Ford was one of them, as he moves more into the philosophical side
that he does a lot of now. And because Ford is Ford, and we kind of, you know,
he is fantastic at everything, he would normally email us on a Monday and be
like, pizza discussion Thursday. So I guess we 3pm had to be there to do that,
and it would be a discussion around certain topics in order to do it. And we got
there, and my other boss, Joe Belowski, who was there, who loved getting into it
as well, we had a discussion on the concept of lineages. And the two of them
were getting into quite a heated scientific argument about what it meant to be a
lineage. And then I stopped them and was like, hang on, what do you define a
lineage as? And Ford was saying that it was like a lineage in a family in terms
of like a lineage of species, genus. These are the lineages that we're defining
out, whereas Joe was defining a lineage as I followed this lineage through my
parents and my grandparents, and it was proper philosophical that they were
arguing and they were not even talking about the same thing at all. But I
learned quite a lot about how to, why taxonomy is there as a useful term, but it
should not be the be all and end all. And then now, June, typically you have
what is a lineage, which is actually now slightly becoming a terminology in
taxonomy, even though it's not properly defined. But lineages we hear a lot
about in terms of people who want to make a clade sound more important than it
is, I think, these days. I think it's a good way of putting it. One thing you
said was we define species on biochems. And actually, I thought it was based on
DNA-DNA hybridization, but it's actually defined on that assay, right? If you
wish to put in a new species into the International Journal of Systemics and
Evolutionary Microbiology, who publish these lists, so there is a list which
comes out every, I think it's year, maybe two or three times a year, I can't
remember, that say, these are the names of the species, and this is what it used
to be called, or this is it, and this is its type strain. And if you want to
publish that, you have to do it in a very specific way. And you can either
publish it in IJSAM, and then it automatically goes into the list because they
do all the approvals to even get the paper done, or you can publish it somewhere
else. And if you have the right amend at the end, and it's been covered, they'll
probably still put it into the list. Now, normally, what that means for a
species is that you have to show biochemically that it is different from other
species that are nearby, and that you use the 16S sequence in order to show that
it is different from other species around it, with the 16S now being the proxy
for that DNA-DNA hybridization. Nobody's doing DNA-DNA hybridizations and
putting all that work in, we can just sequence the 16S. So it's the classic
biochemistry has to be there, and normally some kind of phenotype that's going
to define it, and that strain that you have deposited is classically going to
show that, mixed with the 16S sequence, which has worked as the proxy for that
genetics thing. That's how it currently stands for defining this. More
frequently these days, people who submit to IJSCM are providing genome
sequences. So we have moved, as reviewers and editors, to looking at ANI and
digital DDH, and not just 16S. But certainly, if you don't have the genome
sequence, 16S are similar. So the question of, you need all of the other
biochemical tests, and you need that full description, and then you have a
choice, one of either the genome, ANI calculation, or the 16S, and anything else
you can possibly use, or the DNA hybridization, but no one does that. So the
official regulations say that there should be a 16S tree, but I would say the
vast majority of the reviewers really would now look for a whole genome and an
ANI calculation. So for those who don't know how the ANI works, it is the
average nucleotide identity. You get all the bits of the two genomes you're
comparing, and then you get the percentage of how much of the nucleotides inside
there are shared. And if it's over 95%, that is equivalent approximately to the
16S cutoff of 97%, which is approximately equivalent to the 70% DNA-DNA
hybridization. But ANI definitely works the best of the tools that we have at
the moment for the species level. And I think, even though it's not a
requirement, it's definitely strongly, strongly, strongly recommended if you
wanted to get in. There's a whole lot of weeds in the calculation that we can
dive into for how ANI is calculated by all the different methods you can
calculate. One thing that follows on from this for me is, based on what all
you're saying, and it sounds nice and specific, but then this means that, what
do you do with metagenomic data? And what do you do with genomes assembled from
metagenomic data? Because you then cannot classify them or integrate them into
any such classification as much as people would want to. We are nodding
profusely. Yeah. You certainly can't obtain a phenotypical biochemical
distinction. I mean, this is all about cohesion and exclusion. So for things to
be in the same species, they have to cohere, they have to share properties. And
what we're typically asking for is that they are genomically similar, so 16S,
DDDH, yet distinct from other things which are not the same. So we're looking
for some kind of a transit. So that can be 16S percentage identity, ANI
difference. And for phenotypic and biochemical tests, we're looking for things
within the species to have the same response to those tests, and things that are
not the same species to have a distinct response in at least one of those tests.
You can't do all of those things for genomes that are obtained from mags, and
you can't do it for unculturables either, because you can't do any phenotypic
work on them. But there are a couple of different approaches have been proposed,
but have been rejected by the ICSP, where it was suggested that we should take
the route of allowing type material to include a genome sequence and not the
organism itself in ATCC or some other collection. And that was rejected. So it's
a formalism and a policy decision not to allow mags and unculturables to
participate in this larger scheme that Kam was described of, so much into having
an effective publication and being defined on the basis of both genome and
biochemistry or phenotype. How do you personally feel about, and this is
everybody, how do you personally feel about that? I think we're at a turning
point in taxonomic history, and it will change eventually, even if not now.
Yeah, I would agree. The real clinching point for a lot of people is that we
know categorically that a lot of the microbes that we work with are not
culturable and never will be culturable by themselves. That's just not, in the
worlds of microbiomes that we see now, it's just not possible for a lot of
things. And that is going to then just exclude a lot of things from the
taxonomy. We know that there are a lot of entire phyla that are out there, and
they have been kind of given special status where they say, we know that this is
here, but we're not going to give it an official one, but we don't want to seem
like we're living in the past, that we know that it's there, but without that
cultural, you can't have it. We're just like, that's not, I don't think ever
going to be possible. So I think eventually the genome will become a type
material. And there are efforts to have alternative taxonomies. So what's also
interesting about taxonomy is that there is one true taxonomy, and also there
are 50 taxonomies, all of which are valid. And this is where people get very
confused. If you change something in the official lists, that's what will get
changed at NCBI. That's where people think of when they think of taxonomy. They
go to NCBI, they've got their number, mycobacterium tuberculosis 1773. I can
tell you that's his number. I look at it so much. You put that in and it'll give
you the taxonomy. It'll tell you all the synonyms that it's been before, and
it'll give you that whole tree of where it goes all the way back to the root.
And that is essentially populates that NCBI list. But once something is named,
except for very rare things, that name never goes away. So when they change the
name from Clostridium to whatever it was, everyone will say you have to use this
name, but actually technically you can use whatever name has ever been validly
published for that. You cannot force somebody to use a new name. But in the
reality is everyone wants to have one list and everyone wants to know what
everyone's talking about. So they all want to use the one list. But if that list
is not going to include things that are based on genomes only, people will start
drifting off into other lists. For me, the point of taxonomy is so that I can
say I work on this and you understand exactly what that means. And I don't just
have to like give you a genome accession number in NCBI. I can say I work on my
mycobacterium tuberculosis and you have a general idea of what that means. So it
comes back to the philosophical thing of like, why have a taxonomy if everyone's
not going to use one taxonomy? Just two comments on that real quick. I remember
666 is Vibrio cholerae. And the other one is, at the bottom of NCBI's page, I do
believe they give a disclaimer like this is not the official taxonomy. So I just
wanted to hear a quick comment from you on that. So I didn't know that it was
taken straight from the society. You can ask them to change things without that.
So we found a mistake that there was a mycobacterium species that was in
mycobacterium tuberculosis complex, where the paper specifically said that it is
beside. So someone just put it into the wrong place. I don't know exactly how
the background workings of it works. From my personal experience, what we had
was the mycobacterium genus from 2017 split into five different genera based on
a whole host of things, AAI, ANI, and some other things that are not overly
important. That paper came out, that person put the proper amend records in, it
went into the official list, and it was changed at NCBI straight away. And that
was because it was a genus level paper. It did not take much. Maybe they also
went to NCBI and asked them to change it. But these changes happen very, very
quickly into NCBI. And even though it's not the official taxonomy, people look
at it as official taxonomy. If people say what's the taxonomy, they go to NCBI
and they look it up. That's the reality. This comes back to some extent to the
sort of constrained by history view of taxonomy in the sense of those old names
that have been superseded still persist in the literature. And if you go back to
look at the literature, they're still there. So you won't find that original
reference if you search for the new name. So I've got a paper from 2003 where we
sequenced an organism and it's completely changed its genus now, but the
original paper is still with the original genus assignment. And this touches as
well on the difference between taxonomy and nomenclature. Where if we want to
look at taxonomy as a correct description of evolutionary history and how
organisms radiated, that's one thing. And the labels we put on them is something
different. The label doesn't say what it is. So when we've corrected names, such
as in Lineage A over here, this was genus X, species X. In Lineage B over here,
we said this was genus X, species X. And we've renamed one of them. Both of
those original identifiers will be in the historical record in the literature
and will be in the databases. But only one of them is in inverted commas true.
And if we treat those labels as the truth, as opposed to the underlying data and
what the organism actually is as the truth, then we run the risk of greater
confusion than we already have in taxonomy. We have the same Mycobacterium
tuberculosis and Mycobacterium bovis are exactly the same species. They're just
ecovars of each other, essentially. And then we have Mycobacterium africanum,
which is two lineages.  are not even together on the tree, they have the animal
lineages in between. But when we work on microbacterium tuberculosis lineage 5,
if we don't put in brackets microbacterium africanum afterwards, they'll be
like, people won't know. And I'm like, it's just more confusing. The difficulty
we have there is that then clinicians think, I need to treat this differently
because it's not tuberculosis. I need to find out what treats africanum. So we
have all these like chimpanzee versions and other ones. And then so it's
actually very confusing for vets who tend to be the ones who come, who find new
species or find new lineages of existing species. They're like, oh, I have no
idea how to treat this in this animal. And you're like, it's just TB. You treat
it with mostly the regular things. There will be some differences between
lineages, which we can get into, but sometimes naming can help and sometimes it
can hurt. And then historically, you end up having to say it used to be this,
which used to be this, which used to be this, which used to be this, so that
everyone feels included. I think that's a great example of that tension between
science as a way of understanding and categorizing and classifying and building
from what we know about the world and science as a social activity where we need
to explain to each other what it is we're talking about. And I think this comes
back as well to the definition of species question you asked about, are we
talking about speech in the sense of here's a label for an organism that we can
both talk about and understand? Are we talking about species in the sense of
this is a sensible biological division between this group of organisms and this
group of organisms? And I think we flit between the two of them sometimes One of
the classic examples for me is the E. coli Shigella problem, where basically
it's a specialized inter-aggregated E. coli and it is embedded within the E.
coli species. There is no doubt that it is not something else, but clinically
it's a different thing. Clinically, Shigella causes Shigellosis. It's its own
separate, practically it has to be treated kind of separately. And that's
something, you know, we always make a joke like, ha ha ha Shigella, you mean E.
coli. And it's like, yeah, but there is this sensitivity of like, what are we
talking about? That's always the best example, because clinically you need to
know straight away if it's Shigella, because the way that you treat that is
quite different to the way that you treat E. coli in terms of the severity and
how quickly you kind of want to get to this. So that's really, really important
clinically. So it's about the end, you know, so your use cases, clinician, they
don't care about the genome. They care whether it's Shigella or E. coli, and
they don't, they're not like, well, maybe it should be, it's like, how do I
treat it? Do I treat it this way? Do I treat it that way? Whereas like
technically Salmonella and E. coli and Shigella are one species, but you would
never do that because that doesn't help anybody to talk about their work and it
doesn't help clinicians. So it helps nobody by being facetious. So the example
of that, because I always come back to Mycobacteria, we went through the whole
Mycobacterium genus and we reassigned some of the species into subspecies that
had been separate before. But we have one, which is Mycobacterium marinum,
causes disease in fish and Mycobacterium ulcerans causes very bad ulceration of
the skin in humans. Mycobacterium ulcerans, when you do an ANI, because it only
looks at the bits that are the same between two, which we'll get into later,
these come out as being exactly the same species. But we said, we should not
call these two different things because if someone has a marinum infection,
it'll go away eventually. If they have an ulcerans infection, they could
potentially die from that. You need these labels to be important. I think the E.
coli and Shigella example is a classic because it is such a clear and excellent
example of that tension between the need to name things clearly for societal
benefit and human activity and the need to categorise. And an overdue attention
to nomenclature confuses the two. And maybe we need parallel ways of identifying
what organisms are. It was mentioned earlier, but I don't think specifically
about other ways of classifying genomes than by nomenclature. And I thought you
were going to mention UNCODE or SETCODE, which is a proposed, completely
qualitative of the identity of organisms on the basis of their genomes. Maybe we
do need to separate those things in order to clearly distinguish between the
societal need for classification and our need to classify in terms of
evolutionary history. My book bear will go in the opposite direction, despite me
saying, coming as a clinical taxonomist, I'm like, it's important that we label
things that are useful. But also it's like, I don't need you to publish 5
million papers that each define a new lineage because you don't understand the
difference between a clade on a tree and an actually important definition that
people need to think about. Sometimes it's important that you say that something
is new. And then sometimes you're overreaching in terms of giving things an
importance beyond what is really there. And you see this a lot more in virology
because these things become separated very quickly. So they see lots of
diversity and say, this is a new lineage, but it doesn't actually do anything
different. That's all the time we have for today. I want to thank our guests,
Dr. Leydon Pritchard and Dr. Connor Meehan. We learned a ton about taxonomy. And
so I think you earned your place for the show for next time. We'll go into more
details. Thank you so much for listening to us at home. If you like this
podcast, please subscribe and rate us on iTunes, Spotify, SoundCloud, or the
platform of your choice. Follow us on Twitter at Microbinfy. And if you don't
like this podcast, please don't do anything. This podcast was recorded by the
Microbial Bioinformatics Group. The opinions expressed here are our own and do
not necessarily reflect the views of CDC or the Quadram Institute.