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 is 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. Both
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 U.S. Hello, and welcome to the MicroBinfeed podcast. Today, we
are talking to a special guest about the situation of COVID-19 in Denmark.
Things are changing very quickly, so I should mention that today is the 9th of
March, 2021, and some of what we say may change by the time you hear this.
Joining us today, our special guest is Mads Albertsen, who is a professor MSO at
the Center of Microbial Communities at Aalborg University. Welcome, Mads.
Thanks. I'm happy to be here. Okay. Let's get straight into it. We've talked
before in the past about how getting into COVID was kind of like this phone call
in the middle of the night. How did it start for you? How did you get roped into
this? So basically, I was following it in the start of the year and I took it
seriously, but didn't really know what was coming. But then in the middle of
March, our prime minister closed down Denmark, sent me home from the university,
closed all apps. And then I figured, can we do something? So I basically called
one of my funding buddies, can I start using one of my big grants on this and
all my people? They said yes. Then I called the hospitals I knew, and suddenly I
had contact to the national lab that was doing this. And I said, if you can ship
me samples, I'll sequence whatever you send our way. And that started over
weekends, basically. So Monday, the university labs was opened again. I got in
contact with our rector, got him to open based on COVID work. And five days
later, we had sequenced the first genomes using the ARTIC protocol on nanoform
in Denmark. That's amazing. So you were able to cut through the bureaucracy to
get it done that fast? Yes and no. We're still in the bureaucracy right now, a
year later, I would say. So we were able to sequence within a week, anonymized
samples, test samples we got from the hospital. Then it took two weeks more to
get the legal documents in place so we could actually help out the government
lab. It's called Statens Serum Institute in Denmark. And it's not like we could
do research on it, but we were able to act as like a deliverer of data to them.
That meant we could suddenly handle all the sensitive information at our place.
That took two weeks. In my mind, that was like ages, according to the lawyers,
that was like world records getting that set up in so fast time. They were very
happy. So we got started back then. And yeah, three weeks after we started, we
actually had the first real sample sequenced where we could connect it to
patient information. That's amazing. So, I mean, what were you doing before
this? What was your day job? What was the project originally that you sort of
co-opted for this? So my group normally does metagenomics and we play with
nanopore. So we try to play with the next shiny sequencing instrument we can get
our hands on and see what we can use it for. One of the big projects we had
ongoing is what's called the Microflora Technica. We convinced a private
foundation called the Grundsatz Foundation to give us, that must be like 4
million euros, to sequence 10,000 samples across Denmark to figure out which
microbes are actually present. So it's really like basic science. So we play
with new methods, try to discover as many novel microbes as we can. And those
people, so I have around 10 people on that project, we threw into COVID work
back then. Did you get the publications out from that or did you immediately
have to switch over and that's indefinitely paused? That's indefinitely paused.
Getting back to it now. But yeah, there was a lot of nice high impact papers on
the way, we hope, that hopefully still are there now when we are starting to
slowly get back. Don't worry, everyone in the world has moved over to COVID, so
you know, you're not missing anything. Yeah, I think that's fine. It's been fun
to help out, but of course it's been a big hit to our normal research. And so
how did you actually get started off? You're using MinIONS, did you have any
other sequencers or was it just a room full of MinIONS taped to random
computers? Yeah, so from Twitter, that's how it looks and that's also how it
runs. So we have Illumina also at our place, but we mainly use MinIONS because
it's very flexible for development work when we have to test things. And it's
just a bit more fun to play with Nanopore normally. So that's where we got
started originally. And so what do you do? Do you run 96 at a time on a flow
cell or do you wash flow cells in between or how do you work? So what we do
right now is we run everything in plates. So we have 96 positions, I think we
have 88 samples on the plates. We pool two plates and divide it on two or three
MinIONS at the same time. We do that to make sure if we have one bad flow cells,
we'll still generate enough data. And then we start sequencing in the afternoon
and then we kill it after we have generated 10 gigabases per flow cell or at 6am
the next day. And that makes sure we can actually just deliver data. We use like
one third of the capacity of the flow cell and then we just chug it out. It's
too expensive to start to wash flow cells or do anything else compared to what
everything else costs. Are you not saving up those flow cells for later on for
all of your other normal projects? You can just use your 20 gigs or whatever on
a flow cell? Yes, we are of course washing them and keeping them in freezers so
we can reuse them. And we are also using them quite a lot. So if we have to test
something, we just take a few old flow cells out and test it on that. So that's
really convenient to be able to use it. It's really only emergency where we're
using it to sequence real samples. We had a few times where delivery got stuck
from nanopore or we had a lot of bad flow cells. We then had to go to the
freezer and take those out and they work just as fine as normal flow cells. We
just play it safe normally because it's not the sequencing cost that matters in
these projects we run. It's such a small part of the cost. So you said you do 88
samples on a run. So what does your plate look like? How many controls? How many
positive, negative? So we have four negatives spread in the center of the
plates, divided a bit out. Then we have one positive and we have two blank spots
for measuring concentrations. The nice thing is that we get these plates so we
don't do the RNA extraction. We get these plates from the hospitals or the
national lab and they ship it to us then and then we can just take a plate, just
chug it through without changing anything in the whole process. That's really
convenient from our side that we don't have to touch anything. The room for
concentration measurements is also in there. That is awesome because we get
random plates from places, not even plates, tubes and stuff sometimes from many
different diagnostic platforms. Some of them you don't have CTs, some of them
you have to do extractions, that kind of thing. And my God, we'd love just to
have plates of RNA, it'd be great. Yeah, that's really a privilege in that way
that the national lab in Denmark, they handle all that in an automated fashion.
So in Denmark, we have two test tracks. I think most countries have a healthcare
track based on the hospitals, then we have a community track which is really,
really large scale testing in huge labs on robots. And the flow from there is
quite easy. Everything is done in robots. Positive sample is called, gets re-
extracted in the normal format and shipped to us. But all the handling of
samples from the hospitals is a manual process, but not done by us. That's
really grateful for all the work that the national lab is doing. That's like a
huge work to do that and we get the nice plates. I guess people maybe not
realize just how big of a logistical challenge that is, like that is enormous.
And I know in the UK, that's one of the huge blockers is just simply cherry
picking out the positives from the mountain of negatives. And yeah, that is just
a huge blocker and it has been a rate limiting step for the UK, which hopefully
they're slowly solving, but it is a big challenge, particularly when you're
operating at scale. It's a massive challenge. So fair play to your national lab
for sorting that out. Yeah. I'm a bit curious on that because it sounds like the
national lab already had this infrastructure in place to kind of do this work.
So what were they fielding routinely? Was it just general everything or? Yes. So
in Denmark, from the start of the pandemic, the hospitals were required to ship
out all positive samples to the national labs. So in that way, it was actually
set up that the hospitals already picked out the positive samples. So it's not
the RNA we get from the samples, we actually re-extract the positive samples.
Again, in the beginning, I was very surprised. Why couldn't we just get the RNA?
But that's a logistical nightmare. It's much easier to actually get the samples
itself than do re-extraction again. And so what's the turnaround time from when
someone gets sampled to actually having a sequence? So from the community track,
we have a medium turnaround time of five days. On the healthcare track, it's
around 10 days. That's quite interesting. Why is it slower for healthcare?
Because there's much more logistics involved. So this is done on a lot of
hospitals in Denmark. It has to find the samples and ship it in.  If it's from
the community's track, it's on huge robot farms that's located just next to the
national lab and everything there is just automated. That's interesting because
we have the opposite. Our healthcare, which is pillar one, is very, very fast
and it can be sample one day, sequence the next. And it's a community bit that
takes a bit longer, because samples have to go up and down the UK and God only
knows where else. And then before they actually ever get to a sequencing center.
And so turnaround times can be much longer, like in terms of weeks, like two,
three weeks. Yeah. I see a local turnaround time is also faster. So we've set up
the sequencing at all hospitals in Denmark too, so they can do local sequencing
and they do local outbreaks and there they can do same day turnaround on
samples. And they also do a high priority. But for the national surveillance, it
is around these 10 days, medium turnaround time from the healthcare track. So
how are your clinicians actually using the data for local outbreaks? I actually
don't know much about that part. So we are more involved in the data generation
and then the hospitals are using it for small local outbreaks. But right now in
Denmark, it has really mainly been used as a national surveillance of B.1.1.7
and also monitoring of variants of concern that has then been enhanced contact
tracing on it. I think that's what we envy the UK for quite a bit, using the
data a bit more actively on that part than we are in Denmark right now. Yeah, I
know for outbreaks, it has proven useful, but it requires someone like one of us
to actually interpret the data for them, because while they can look at
lineages, it doesn't really tell them very much. It's when you actually have to
dig in and say, this over here is at luster and that's not at luster, or you've
two outbreaks here, or it's kind of one thing is split and now your staff are
infecting staff and patients are independently infecting patients. It's proven
to be quite useful, but you do really need that genomic epidemiology and those
experts on hand to interpret the data. Yeah, we have a few of them in Denmark at
the national lab, but as everyone else, they are overburdened with work. So I
think they are in short supply. People can actually do that. And that's the main
bottleneck actually for using the data. That's having more people who are very
good at this, interacting with other healthcare professionals. So what do you
think the future is going to be like now in Denmark? Now you've gotten
sequencing into the clinic, which is what we've all been trying to do for years.
Do you think it's going to continue? Is it going to just kind of fizzle out? I
think it has really opened some doors and shown people what's actually possible
today with sequencing capacity. That it starts to be too expensive not to
sequence the cases. And we have the technology today to simply sequence
everything. And I think we're going to move very fast on that. Denmark has
already been okay at it. But I think now the local hospitals also start to see
the value of doing this. And then the main bottleneck is, do we then have enough
people who are actually good at this to make sure the data gets used? Because I
can see, you know, in the future, people try and apply the same methods for flu
and for just any viruses and bacterial outbreaks, you know, just in general, you
know, as a matter of course, and that will be quite interesting. But then, of
course, you can't have amplicon sequencing for every different virus and
whatnot. It just wouldn't make sense. So it would really be down to doing
metagenomics and microbial communities, which you are more familiar with than
most. Do you think that would be a reasonable thing for hospitals to have, you
know, metagenomic sequencing, just sequence a sample, and then you get an
answer? Yeah, I think that that's given the cost of sequencing today, I think
it's relatively cheap to actually do. And most people have the sequencing
capacity to do it. So I think that could be something that moves into routine.
But of course, what I've learned last year, that you need to have really robust
protocols and procedures at the hospitals, and that takes some time to develop.
We just develop something, and then we let others make it into some robust
workflow. And that takes some time to make sure that's really integrated. So
what I want to know is, you started off with the ARCTIC protocol, but then you
tweaked it slightly. How did you tweak it? And why did you make the changes? So
we started back in March. And our main problem was there, if I did back off the
envelope calculations, we had 24 barcodes at that time, I would need like 100
men in the lab to actually do these protocols. So the one thing we wanted to
change was barcoding. So we switched to barcoding in the PCR, which is very easy
with nanobar. I don't know why more people don't do that themselves. So that was
the first thing we changed. And then at the same time, we wanted to use it as
the capacity of the nanopore, and that's a bit better if you have longer
amplicons. So we dialed it up to around 1200 bases to be able to push 30
gigabases to flow at least. So that was the main changes, barcoding at the PCR
step instead to allow for, I think we made 500 different barcodes or something
like that back then. Did you use the nanopore PCR barcodes? No, we just made our
own, put in 10 random nucleotides, and you have your barcode, and then you call
it. So it's really easy to do. So I'm curious with the long amplicons, what
we've noticed with some of the variants of concern is that the mutations may
take some primary targets out of, you know, they don't work and you won't get
any coverage in that area immediately downstream. Have you been seeing that on
to your protocol? So we actually, so we use the Arctic primers too, we just use
one third of the primers basically. And we do also see amplicon dropouts from
time to time, not in the variants of concern currently, but if we see an
amplicon dropout, it's going to have a larger effect because we miss a larger
part of the genome. But on the other hand, it's also a smaller part of the
genome sensitive or having mutations because you have fewer primers. Yeah,
that's true. So there is a deletion that is causing some issues and they are
bringing out an update to Arctic, just one extra primer to fix it. Yeah. So what
we've done so far is we do rebalancing. So we add in more primers of those that
seem to not work that nice. We had some issues with primer based dropouts back
in November when we have all the mink variations, but they disappeared again and
then our problems disappeared. It's really difficult to fix these problems
because it depends on the current balance that actually circulates in general. I
think the current Arctic primer seems to work quite nice still. So I've heard
that the reason why Arctic goes for short amplicon lengths of about 400 is
because it works better on high CT and degraded samples. What kind of effect do
you see by increasing to 1200 or 1400? I think we compared around a hundred
samples originally with the original Arctic compared to what we do. It was very
similar results and that's basically what we've done so far. And you can see if
we compare with what people do in the literature, we are at the same success
rate more or less. We were also concerned, especially in the start, that we
might miss a lot of the high CT samples, but it doesn't seem to be a big
problem. I would assume that definitely smaller amplicons are better, but in
practice it doesn't seem to have as big effect as I thought. Do you think it's
because in Denmark you're doing the re-extractions much closer to the
sequencing? I know with us, the RNA can be quite degraded sometimes and we just
have to use the diagnostic, like the waste from the diagnostics and whatever
comes out of the machine, that's what we have to use. And it might be sitting at
room temperature for a while, there might be a freeze-thaw cycle, but actually
you guys have a much more controlled way of getting those samples and maybe
that's why you can actually go for these longer amplicons. Yeah, that's a good
point. That could actually be one of the reasons why it works quite nice. And do
you have a publication on this new method? We made a protocols.io back in April,
we just haven't had time to do anything. It's a bit crazy. We had hoped to have
like 10 major papers now on everything, but just haven't had time to do much.
We're starting now to get a bit back to writing some of the things up we've
done, so writing a bit more on the performance on the method and also what we've
seen in Denmark. But it's first now and let's hope we don't get any new
variants. So we thought back in November that we should start to now write
things up, then we had all the mink variants in Denmark, then we started to get
a bit relaxed in December, then we have the B.1.1.7 and yeah. So whenever I
think we have to relax, we get a new variant, we have to... So speaking of mink
variants, we've just seen an N501T pop up in Zimbabwe, which is quite
interesting. Yeah, yeah. I would say long Amplicon protocols out there that also
seem to work nicely. This is just what we've started doing and what we are still
more or less doing. So what kind of scale are you guys operating at now? So
right now we're doing 1000 samples a day, but in theory we can do around 2500
samples a day. But we have to push. And we do two days in the lab per batch. So
we start doing PCR and cDNA on day one, and then we do pooling, normalization
and nanopore prep on day two. Do you guys work weekends as well? No, not unless
it's really urgent. So we did in the start of January, there we said you ship
everything, we'll just sequence it. And I think we did a couple of weeks with
10,000 samples a week back then. That's more than most countries have done, you
know, for the entire pandemic. So this is, you know, vastly bigger than Denmark.
And you do it in a week. Yeah, it's quite nice. But also, as we talked about
before, it really also comes back to how it's set up in Denmark on a national
level with how the testing facilities and the sample prep. So if you got samples
in a nice 96 world format, I think you can also push more samples. Now, just as
a comment, like this seems to be a common story that anybody who's really he had
successes on the back of long term infrastructure, long term investment in
knowledge. Just a hint to anyone in policymaking out there. Well, you mean you
can't just build this overnight and like America's trying to do? Yeah, you can't
just pile of money at us. You can't just throw money at it. No, that seems to be
the issue. I was just gonna ask, was your Christmas cancelled? Like our
Christmas?  was over here. Our holidays were very reduced at least. So people
were tired after our November where we really scaled up to sequence the mink
related cases. So we had plans for going for a thousand samples a week instead
of five thousand, but we powered through instead. We still had Christmas, still
had a few days of holidays, but other than that we had a team in processing
samples. And the data we generated around Christmas and New Year's, that was
actually the data that was used to enforce stricter lockdowns in Denmark. So I
think we also had an impact that we really stepped up and generated enough data
that the politicians there take decisions based on that. And that's a really
good impact, that's science in action, which must feel really good for you guys.
You know, you're doing a bit of work and then you're seeing the effect
immediately. It's not like you're writing a publication and waiting, you know, a
few years to see will someone hopefully pick it up or cite it. Yeah, it's a bit
strange that we actually see that this has an impact on the society as we do it.
And of course the team is very proud that we are helping to do this, but it's
also very special that so much emphasis is now put on these data. They actually,
that we have a government that's quite brave enough to try to steer some of the
country, basically. I mean, some of us said that over Christmas we were getting
blamed for the whole thing. Do you really want to be known as the scientist who
stole Christmas? But that's also how it is in Denmark. So we also start to have
more people making a lot of noise and are not happy. And I also get funny,
strange emails from strange people saying that this is our responsibility and
there's no virus. But I guess this is what we have to live with. So based off
what you're saying, it sounds like most of the organization goes to the National
Reference Lab. And is that also applied to metadata, contextual data around the
samples as well? How does that flow through from the hospital to you? Yes, so
everything is going through the National Lab. It also keeps the national
surveillance data for Denmark. That means that every single sample can actually
be connected directly to a person. And that data can also be connected directly
to the health record that can be pulled by the National Lab. So everything in
Denmark is really directly connected to the data through the National Lab. And
by coincidence, more or less, it's actually my group that then connects the
genome data to all the healthcare-based data on our servers. And who manages the
uploads to GISAID and the ENA in place of that? That was my fantastic PhD
student, Thomas Using-Mikaelsen. He does all the uploads until now without any
errors, more or less it seems. At least we have like a perfect record of dates
as I've been following on Twitter compared to many other countries. I see art is
doing up these beautiful graphs where you're showing, well, kind of shaming
countries. You're very happy about them in Denmark. I know in the UK's case,
sometimes you just can't get dates because sometimes samples would be slipped
through the system. Or actually, sometimes you get the data sample collection
with a person's date of birth. And it's like, well, probably not. We get a lot
of strange samples in Denmark too. And we have around 50,000 sequences that's
uploaded. We've sequenced 100,000 samples, I think around 20,000 of them,
multiple sequences from the same person or things that, by one or the other way,
couldn't be related to a proper sample. So we don't upload it. So we have a big
filter in front that makes sure that the data we upload from Denmark is also
quite representative of the pandemic. Awesome. And what kind of data are you
allowed to share? Not much right now. So we are allowed to share data with a
date that binned to a current week and the region where it was taken. And I'm
not sure we are really allowed to do that by the lawyers. That has really been
an ongoing discussion. How much are you actually allowed to share? And that's
why I tried to say before that we are still discussing with the lawyers a year
later on what we can upload and what we can do. And also how can we share, for
example, data just back to the hospitals. That's actually also difficult because
they take samples at the hospitals, but they also have a lot of community spread
that's tested in the community track. But we cannot share those data with them.
So there's still, I would say, a lot of strange legal things that needs to be
sorted out to be able to use these data effectively. And I think that's
hopefully for the next next pandemic. And that's hopefully sorted. It really
took us by surprise. I think this time how difficult these things are. That is
quite disappointing, because it is quite useful seeing the community data and
the hospital data in context, because you can tell quite easily if this
particular outbreak is related to community or if it's just hospital associated.
You can see it advance nearly. Yeah, in Denmark we can also do it, but it needs
to be in collaboration with the national lab then, because they're the ones with
the correct level of access to data to be able to do it. But that's really a
bottleneck that you need to then have like 50 more people employed at the
national lab to just help with inquiries from the hospital. So I think that's
one of the things that needs to be sorted, at least in the future. And we are
working on it. It just takes ages when you have lawyers involved. So what are
the next steps for your group as vaccination is rolled out? So that's getting
out again of this basically. So when we started back in March, I set a hard
deadline, say, we want to be out 1st of January again. And that point to the
hospitals and the state lab needs to have taken over. Then we have a lot of new
variants coming out. But now the deadline is summer, we want to be out. And
that's also what we are setting up now. So the national lab is building
sequencing capacity, and the hospitals also building capacity. We now have some
of the big hospitals starting to see a sequence a much larger share. And that's
what we are using the next months for figuring out how can we actually make sure
that we can improve the surveillance actually. So I think bringing the
sequencing even closer to the samples is going to improve the turnaround time in
a good way. So hopefully we can land that nicely the next three months and help
with that transition too. And you can have some summer holidays then? Yeah, I
would like to be out before summer holidays this year. But of course, if it goes
wrong again, if we get new variants because of the vaccines are popping up, then
we of course here to help. But we hope like everyone else that this starts to be
over by summer. And that we all can get back a bit more to normal life again. So
who's taking over at SSI actually? That's also a bit undecided now which group
is actually going to run it and if it's going to be together with tests in
Denmark, so the big testing facility. And that's what we are figuring out now.
SSI is figuring out how it's really going to set up. But once you have programs
at this size, this is actually a quite big program. There's also starts to be a
lot of political interest and people with big egos that wants to figure out how
this is done in a proper way. So it's a bit more complex process getting it set
up. But again, we are here to help and so that's really in place and hopefully
it's going to be set up in a few months. So what are some advice for anyone, any
country or any lab that wants to get started in this space? You know, just a
couple of pointers. So I think one of the important point is to be inclusive, to
make sure that you acknowledge the hospitals and all the people actually
involved in this. Sequencing is one part, but you need to really have excellent
collaboration and the whole like food chain from a sample is taken. So you
actually get it acting on the data. I think that's been one of the big lessons
in Denmark that we have really a good healthcare system in place already where
these things are relatively easy to do. And don't neglect how much work you
actually need to put in to get it to work. Then I think the other thing is make
sure you solve all the problems that comes your way. I think one of the reasons
why we've been successful in Denmark is that I have an awesome team behind me
that solved a thousand problems that occur in projects like this. Just from
primers missing in the lab to not being able to order reagents to figuring out
how to buy stuff in Singapore or stuff like that because it's just not
available. It's a thousand problems that's going to occur and you need to put on
the hat saying I'm going to forget that it's a problem and they just solve it.
On that inspirational words from for me, Mads, that'll be it for today. So
that's all the time we have for and yeah, I want to thank Mads for joining us
today. Very interesting. And thanks to everyone for listening. Thanks. Thank you
all so much for listening to us at home. If you like this podcast, please
subscribe and like us on iTunes, Spotify, SoundCloud or the platform of your
choice. And if you don't like this podcast, please don't do anything. This
podcast was recorded by the Microbial Bioinformatics Group and edited by Nick
Waters. The opinions expressed here are our own and do not necessarily reflect
the views of CDC or the Quadrant Institute.