Hello, and thank you for listening to the Microbid Key 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 is so much information we all
know from working in the field, but nobody really writes it down. There's 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 Professor Andrew Page. Nabil is a
senior bioinformatician at the Center for Genomic Pathogen Surveillance,
University of Oxford. And Andrew is the Director of Technical Innovation for
Theogen in Cambridge, UK. I am Dr. Lee Katz, and I am a senior bioinformatician
at Centers for Disease Control and Prevention in Atlanta in the United States.
Hello, welcome to the Microbid Key Podcast. We've got Costas Konstantinidis.
I've known Costas for almost 20 years now, I think, actually, from Georgia Tech.
So we have a good relationship. I'm going to put out my bias there right away.
It's been a pleasure knowing him. He's one of the most friendly professors I've
ever known. It also turns out that he is one of the originators for ANI and a
whole bunch of metagenomics stuff that we'd like to talk about too, but maybe we
can get into ANI first. But welcome, Costas. Hi, Lee. Hi, Andrew. Very nice to
be here. And good to see you again. So can I ask, we have never met, unlike you
and Lee. So what's your background and how did you get into the field of
metagenomics and ANI and this murky world of environmental metagenomics? Yeah,
so I'm originally from Greece, as you can tell from my very Greek name, which
Lee pronounced it very well. I did my undergraduate in Aristotle University in
Thessaloniki, studying, actually, plant protection, agricultural sciences. And
then towards the end of my degree, I wanted to do, I decided I wanted to do
graduate studies, but my family could not afford it. And so I almost, you know,
signed up to go to the army because for the Greeks, you have to serve the army
for two years. And then, you know, the last month, basically, before graduation,
one professor came to me and said that there is a nice fellowship for somebody
to study in the US for four years, the Vuyukos Fellowship. And so I applied for
it and I was granted the fellowship. And that's how my dream became true. And
then, you know, I had my money, my own money from the fellowship. So it was easy
to start with almost any professor because I was a free ride for them. And so I
asked a few of the previous Vuyukos fellows who to work with. I wanted to study
soil microbiology. I had realized that we know a lot about plants, how to
protect them, how to grow them, fertilizers. But the soil microbiology was an
unknown thing. That was for my undergraduate degree, the summary. And so I wrote
to Jim Pizzi and a couple other people and Jim accepted me. And that's how I
came to the Michigan State to do my PhD in East Lansing. And then how they and I
started, etc. This is also an interesting kind of story. So this is now 2000
when I started my PhD. So that's the very early of the genomics. If you
remember, the first microbial genome was sequenced in 96. The first E. coli in
97. 2000, 2001, we had the first three E. coli to do comparative genomics. So I
realized that, you know, this is an upcoming field. And actually, my project
was, I needed to size the genome of some Bulkoderea species using pulse fields
and electrophoresis, which probably is obsolete by now. But basically, it was a
nightmare, you know, to try to get these large chromosomes sorted on the PFGE.
And so that's when I started talking to Jim and said, Jim, you know, sequencing
is coming up. Maybe that's another way to do it. And of course, Jim, you know,
he's very, he was very much into it from the beginning. He sent me actually to
the Institute for Genome Research, the TIGER. TIGER doesn't exist anymore. But
TIGER did some of the very first microbial genomes. And so I went there for an
internship over the summer. That's how I learned how to do the sequencing
analysis, etc. And then I came back to Michigan State and start working with
genomics from then. That's how I ended up working on the genomics field. And
then they and I was, they and I was an idea we developed during my PhD. And it
was from the need to be able to distinguish between these closely related
Burkholderia strains that we had in the lab. And I was trying to size the
genome. And, you know, I try a couple of the original things, like, for example,
the 16S identity that was and still is a big thing, but it didn't resolve them
well. And so at that time, a visiting postdoc, Johan Goris from Belgium, from
the taxonomy group in was visiting Michigan State. And Johan was very good at
doing DNA, DNA hybridizations. And then talking with Johan, I don't know, over
coffee one day, we said that, you know, the DNA, DNA hybridization is such a
great method that has been very influential, etc. But it would be nice if we can
see what correlates best with the genome so we can distinguish closely related
strains. And that's how we came up with the A&I idea. Basically, we try a lot of
the things, a lot of metrics from the genomes that we thought were useful back
then, like, for example, GC content, the gene content, and A&I was the best
predictor, the best metric that correlated with the DNA, DNA hybridization.
That's how we started the DNA, the A&I work. That's awesome. So all my PhD was
basically on the isolates from soils. And, you know, I realized that, OK,
isolates are good, but it's very important to see in what context they leave the
metagenome. And the metagenome was, you know, that's when it started, 2004,
2005. That's where the first metagenomics papers came out. And so I realized
that I wanted to study also what happens in the whole community. And that's how
I decided to do my postdoc in metagenomics. And there were only three or four
labs back then. And I wrote to Ed DeLong at MIT. He was at MIT at that point.
And that's how I ended up there and specialized on metagenomics to working in
the deep sea first with Ed and then expanding on other environments. So that's
more or less my story in the early, early years. I have so many questions.
That's such a good introduction, I think. So can I ask, right? ANI is used
basically to define species these days with genomics, but what do you define a
species as? It's controversial, by the way. Right, right. So as you may have
seen in the recent publications we have, what we really observed later on was
when you take all these genomes and you compare against themselves, there seems
to be an area of ANI that is very low frequency. And that's between 80%, 85% and
95% ANI. So basically, if you take all the genomes from NCBI or EBI, et cetera,
and you compare them, there is a clear gap of ANI around 85% and 95%. And we saw
the same thing in natural populations with metagenomes. So I am convinced by now
that this 95% ANI is a good proxy for species because basically the natural
pattern of diversity shows that there is a gap around 95%. And sometimes it's
not 95%. Sometimes it's 98% or 92%. The SAR11, the most abundant organisms in
the sea, they are usually more diverse within the species. But it looks like
there is this natural gap. And not only us, others have observed it too. And so
I am convinced right now that that is a good approximation for species. That's
based on the natural patterns of diversity. And what I want to say that is more
recent of the press, we have a couple of papers that are coming out now in MBI
and Nature Communications that we saw that there is an ANI gap within species
that we didn't observe before. And I can explain to you why. But the bottom line
is, you know, look them up on the internet. They should be available in a couple
days, weeks. And they show that there is an ANI gap within species. And I think
this is useful to define the subspecies units. For example, genome of us, that's
how we call them. We decided to call them and also strains. So basically, I
think the natural patterns show that there are some gaps that we can use to
define species and they seem to be correlated well with the organisms within the
class to be more uniform and distinct from other organisms that are in the same
sample or in the same habitat. Right. So I do a lot of work, obviously, on my
genomics as well. And I find it quite difficult because you look at NCBI, right,
where a lot of people say this genome is this species and are classically
defined. They may be defined 50, 100 years ago based on phenotypic
characteristics. And then you have, say, GTDB, which is.  going off basically A
and I, and they're saying, actually, that's not a species, or they're split in
these different ways. And they're coming up with all these novel ways of
chopping up the species boundaries. And so what, how do you think the future is
going to look, certainly, let's say NCBI and the kind of gold standard
taxonomies, are we going to have to redefine all of these classically defined
species based on what genomic information we know? So, for example, Shigella and
E. coli is a classic example, you know, or, say, Mycobacterium tuberculosis
versus Mycobacterium bovis, which are 99% identical, you know, should we go and
redo everything? Or should we just kind of say, okay, we know this works? Right,
so the difference or the discrepancy has to do with traditional, like, for
example, if I take the Shigella example, you mentioned, so Shigella apparently
has a very different antigen than E. coli, and the doctors saw that, and they
decided to name it quite differently. And because that antigen is so important
for the clinical diagnostics, it is maintained until today. But today, you see
that actually, it's very closely related to the traditional E. coli, somebody
can say it's a genome of E. coli. And so I think looking forward the future, I
think it's important to realize that, that genomically they are so similar, and
understand also where the differences are coming from, which I think we have by
now, we have understood that. And if you are asking me, you know, how to proceed
in the future, I think my view would be, yes, I think it will help to rename
some of these, but also keep the keep the diagnostic thing, which is important
for the clinic somehow. So maybe as a clear genome of a subspecies of E. coli,
but with a clear defined, you know, phenotype. And so the discrepancy is mostly
because of these traditional names that still exist that were based on different
methods. And I believe what has been named the last decade or so, it's quite
consistent with the genomics that there's not so much discrepancy. So it's
mostly these, you know, organisms of medical importance that have been named
historically that they show the discrepancy. And I know that the medical field
is very hesitant to do changes. I think scientifically, it makes sense, but it
also makes sense to somehow keep, you know, these diagnostic phenotypes that
distinguish these groups of organisms. So, so they are useful for the clinic and
so on. I kind of have the opposite problem because, well, I have both, I have
both problems, right? So, so like Shigella, it's an E. coli basically on the, on
the genomics, but, but then there's, there's the other side of the problem too.
Like when we look at something like Salmonella, it's a species, but like our lab
has found that the species barrier is like a 92% ANI, I think. This paper is
actually going to be out by the time we put this recording online, I think. But,
so, so I can't remember the, the exact ANI threshold, but like Salmonella is so
diverse. And then you can look inside of it, Salmonella enterica, and those
subspecies barriers that you were talking about, we see lots of those, they're
like distinct places, different thresholds to define like the different
subspecies within it too. So it's like clinically, like we want to keep
Salmonella enterica what it is. We don't want to change it, but, but genomics,
like in the, in the phylogeny are showing that, that there are distinct lineages
or subspecies really of Salmonella. So Lee, I, I am not very familiar with the
Salmonella, so I don't want to say more about it, but I guess what I'm trying to
say is that I think we need to look at the natural diversity of these organisms
and see what the patterns are and go with them, like, you know, to name them
species or subspecies. I am convinced by now that in nature, they form these
clusters that we can actually recognize with the genomic tools. And ANI is just
one of the two other approaches also give very similar results generally. And I
think I like that because it doesn't have the limitations of the older methods
that lack that resolution. So I think going forward, I think things look pretty
good in my view that, you know, we can reliably distinguish the clades and
subclades with the genomics tools. And the other thing I wanted to say is that,
and I hope people realize that, you know, one number doesn't apply to
everything. This is biology, this is organisms, but they have very distinct
ecologies, very distinct lifestyles. So we shouldn't expect that, you know, 95
ANI will apply to all of them. And that's why I think it's important to look at
the diversity in the genome we study and see what the patterns are and keep in
mind that maybe for different organisms, the gaps in diversity will be in
different regions of ANI or another metric. Absolutely. It's incredible how much
people want to conflate the genomics versus like the clinical outcomes of like
what we want to finally call the species. So like, like Shigella is a great
example. I like that example. So it's like, there is a distinct clinical outcome
with it. It needs to be defined as different species, but on the genomics level,
like it's incredibly similar to E. coli. I think it's tough to figure that out.
Yeah, I agree. Except if you look at the diversity of Shigella genomes, they are
separated from what is called E. coli. There is an ANI gap there. They make a
distinct clade on ANI or another tree. So yes, genomically very similar. If I
remember well, the ANIs are around 96, but there is a clear distinction. There's
a small gap there that distinguishes them from E. coli. So I think if people
want to name it a different species to keep the medical importance, I think the
gap justifies that. Just maybe they need to change the genus name to be a
Ceresia, not Shigella that makes somebody believe that it's so different. No,
it's not. It's actually very similar. But there is a gap, and that's what I'm
trying to say, that sometimes you need to look at the data and decide where the
gap is, see where the gap is, and maybe that's your clade. And whether to name
it species or not, it will depend on aspects like this, you know, the medical
importance. And I'm okay if they want to name it, you know, I don't know, E.
ceresia flexneri and so on. And I think GDDB, that's what it does. I remember
they were doing that a few months ago. Yeah, it causes all sorts of problems,
because like, I use GDDB as like the taxonomy for building some databases. And
of course, then our public health customers are like, hang on a second here now.
No, we want this called Shigella, not something else. So, you know, it's that
difference between the academic taxonomies and the practical clinical phenotypic
taxonomies that have been around for a long time. I want to say something about
this. So ASM organized a session in the micro conference. Was it last year or
the year before? Anyhow, they invited me to represent maybe the ecology and
taxonomy, but it was in a clinical session, like it was more the clinical
microbiologist. Bottom line is, we are talking different languages, we are in
different worlds. They even have their own journal to do the naming, which is
isolated from the rest. And I think that doesn't help. And I think we need to
talk more to each other. And maybe ASM has a big role to play there. Because I
think there is common ground. And I think we all benefit if we communicate more
between the clinical field and the more traditional, I would say, taxonomy and
ecology. I think there's not so much talking there. And I think that's a
problem. So I noticed one of your publications is on SeekCode. And we actually
interviewed some other people who are involved in that project as well. So I'm
wondering, what do you think of that for naming new novel species and genus and
whatever? Point of information just for you, Andrew, really quick is that
Miguel, the person we interviewed, was in Costas' lab. Oh, actually. OK, great.
So this is Miguel Rodriguez. He's a former student and postdoc. And he's an
assistant professor in Innsbruck, Austria. He's a great guy and a great
scientist. So SeekCode is another, to make things more complicated, there is
another issue. And that has to do with the traditional taxonomy. So the ICNP,
that basically rules the names of the microbes. The problem of ICNP and why
SeekCode emerged is that basically, for uncultivated taxa, they don't give a
stable name. They don't keep track of them. Now, recently, with SeekCode, they
start doing that. And I think that's in a good direction. But basically, to make
things simple, if I have an uncultivated taxon, and I want to give it a name, I
have to name it Candida to something. But then if somebody isolates that
organism, they can deposit the isolate in the cultural collections.  and propose
a completely different name, whatever they like. So overwrite the name I gave to
my candidate that was taxa. So that's what I mean, the traditional taxonomy, the
ICMP, the National Code for Non-Negligent Prokaryotes, that's what it stands
for, doesn't give priority or even regulate the uncultivated taxa. And so me and
many others, among the founding members of C-Code, propose that we need to
recognize the uncultivated taxa. They are real things. Sometimes they are even
more important for what we do. And the big problem was that in order to give a
stable name, you need to have a culture and deposit it to two cultural
collections. And so that's not possible for the uncultivated taxa. So we propose
instead, especially Barney Whitman from University of Georgia, he proposed that
the genome sequence can serve as the type material to go with that name, like
instead of depositing a culture, depositing a genome. And so the ICMP voted on
that and rejected that proposal for their own reasons. And that's a big
controversy there. I'm not sure if we're gonna go there. It might take us a lot
of time, but the bottom line is that what triggers C-Code, and now C-Code is up
and running, and people that work with uncultivated taxa, they can name them
under the C-Code and give them stable names, as long as they have a good genome,
like a high quality genome. So I do think C-Code is the future because in this
era we are with all these genomes and our effort to describe the diversity that
is in the environment, I think the C-Code is the only one that can scale up
basically. So, but it's, right now it's antagonistic to the traditional ICMP.
And again, I hope there we talk more to each other and we have one system, one
united system. In the future, right now we have two systems and that's not
ideal, I think. So this has been an awesome discussion with Costas from Georgia
Tech and from a lot of other places, as we've learned. And this was one of the
most natural, most engaging conversations. And I really do appreciate that. I
mean, you guys are great hosts. So I think kudos to you as well. Thank you so
much for listening to our podcast. If you liked this podcast, please subscribe
and rate us on iTunes, Spotify, SoundCloud, or the platform of your choice.
Follow us on Twitter at MicroBinfy. 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 Center for Genomic Pathogen
Surveillance.