Sequencing samples - COG-UK's key ingredient

Sample types and format

Diagnostic samples taken from patients with COVID-19 lay at the heart of COG-UK's sequencing operation. Essentially, two types of sample were potentially available. The first was a throat or nasopharyngeal swab taken from a person which had been stored in individual tubes with a viral transport medium, a solution to preserve the specimen. Although readily accessible, especially from hospitals, such samples had the downside that they contained live SARS-CoV-2 virus which, for health and safety reasons, required deactivation before they could be sequenced. Additionally, they could not be sequenced without a number of other steps.

Peijun Zhang, a research technician who managed the COG-UK operation at the University of Sheffield, explains, 'The first step is to extract viral RNA from the swab sample' and then we will do cDNA [complementary DNA] synthesis which turns RNA into DNA. Then after cDNA is synthesised, PCR [a polymerase chain reaction] is used to amplify regions of the genetic material. The PCR product can then be used in library prep for the sequencing, and then once the library is prepped, we can load into the sequencer (Zhang transcript).

How the extraction process was performed is related by Ian Goodfellow. His team at the University of Cambridge undertook some extractions of the virus using so-called 'manual' methods from clinical samples at the beginning of the pandemic. Not having an automated extraction instrument at the start, the work was very time-consuming and repetitive. Goodfellow describes the work 'as absolutely brutal for the guys in the lab. It was hundreds of individual columns being spun in a centrifuge at a time. It was really, really boring and tiring work because it's the same thing day in, day out'. Eventually he acquired a semi-automated extraction platform but there was 'still quite a lot of manual work involved' (Goodfellow transcript).

Figure 9.1: Photograph of individual tubes with RNA extracts ready for sequencing at Portsmouth Hospitals University NHS Trust's Research Laboratory. Credit: Sharon Glaysher.

What was much easier to handle was the post-PCR product where the RNA had already been extracted. The advantage of this second type of sample was that such extracts were commonly left over after a diagnostic test had been performed. It was also much safer than a swab because the extract did not contain any live virus and needed less preparation to be sequenced. For Dr Catherine Ludden, who was at the front end of sourcing samples for COG-UK sites, the 'dream sample' was 'a full [96 well microwell] plate of RNA, or even just any plate that has RNA, but no negatives, because then you can just pick and get them sequenced. But that very rarely came from hospitals, it was usually just here's the positive sample from hospitals' (Ludden and Blane transcript).

Most of the RNA samples supplied in microtitre plates came from the pilar 2 community testing carried out in the Lighthouse Labs. The bulk of these went to the Sanger Institute for sequencing, but sometimes they were also distributed to academic laboratories when the Sanger Institute was overloaded. The reason that the Sanger Institute concentrated on sequencing RNA extracts supplied in microplates was because they were more amenable to high throughput. Ludden explains, 'you can't sequence as quickly and as efficiently if you get individual tubes.' She says the plates worked out really well for the Sanger Institute because it meant 'they just had one process to do and go with it.' By contrast, other COG-UK labs 'could be sent two different sample types in the same week, and they had to organise the workload to be able to do both' (Ludden and Blane transcript).

Figure 9.2: Photograph of 96 microwell plates used for sequencing COVID-19 samples at the Sanger Institute. Credit: Dan Ross, Wellcome Sanger Institute. A microwell plate is a flat plate with multiple 'wells' that function as small test tubes. One of the first microplates was designed in Hungary by Dr Gyula Takatsy in 1951 during an influenza epidemic which led to a shortage of laboratory equipment including test tubes and pipettes. His innovation arose out of a need to carry out cost-effective batches of blood tests (Mishra; Jaquith). Ludden believes that 'getting anything into plates is the best way rather than individual samples, as long as it's done in a quality assured way. You don't want someone just putting samples into plates and then mislabelling the plate. You need to make sure that you've got a trained scientist who knows what they're putting the sample in the right well, and that it's recorded correctly' (Ludden and Blane transcript).

The format that the samples arrived in had a knock-on effect in terms of the speed and efficiency of sequencing. Ludden points out this affected capacity, which she says was 'not black or white.' She elaborates: 'Your capacity is dependent on what sample type you give them. They can do a lot more sequencing if you give them [multiwell] plates of RNA than if you give them individual tubes. There were a lot of variables to consider. You need a different infrastructure for the different sample types. If you are just doing plates, a lot of it is done with robotics, but if it is individual tubes, it is quicker to get someone pipetting it' (Ludden and Blane transcript).

Condition of samples on arrival

Whether the samples arrived in individual tubes or in plates, one of the biggest challenges at the outset was how they were packaged and labelled. Some idea of the complexities involved is captured in the interview with Dr Sharon Glaysher who supported the COG-UK effort at Portsmouth Hospitals University NHS Trust's Research Laboratory. She recalls 'when the samples were first shipped, we just had … tubes freely floating in a box of dry ice. We had to rummage around inside of them trying to make sure we found all the samples… A lot of the time data didn't come with the samples. So we had to read the handwritten labels on the side of tubes and try to find what they correspond to. Or when they would send a plate and say “this is what we've got” and it does not correspond to what they had told us before. It was a steep learning curve for them, as well as us, trying to coordinate that' (Glaysher transcript).

In normal circumstances, Glaysher and her team were used to visiting different sites before setting up research collaborations 'where we could run through feasibility checks and transfer protocols.' But in this case, 'because of time constraints and everything moving so quickly', they instead 'relied heavily on video conferencing, software and phone calls.' She says they put in a lot of effort to get hospitals to send 'the samples in a way that worked with our local workflow, because a lot of them would want to send individual tubes, and at that point we were scaling up so we wanted to move from working on individual sample aliquots to working in [multiwell assay] plates. We tried to get everyone to send samples in the same way to help with workflow and throughput' (Glaysher transcript).

Figure 9.3: COVID-19 samples stored in freezer at Portsmouth Hospitals University NHS Trust's Research Laboratory. Credit: Sharon Glaysher. Angie Beckett, who worked alongside Glaysher in Portsmouth, recalls 'We asked people to give us barcoded plates if they could, and to include the unique barcode identifier on the metadata spreadsheet so that we could link the plate to sample records.' But not all of the sites that sent us plates had barcoded plates, in which case when we received them, we would give them a unique ID and label them ourselves with the date and the site' (Robson and Beckett transcript).

The effort to educate laboratories was not confined to Portsmouth. Joanna Watkins, the deputy head of the Pathogen Genomics Unit (PenGu) at Public Health Wales and her colleagues also developed analogous strategies because the condition in which the samples arrived was a real 'bugbear' for them. She recalls the samples would 'come in tubes' and frequently came without being organised. 'We tried to get them to send Excel spreadsheets of the samples, so it was easier for us to transfer into our recording databases. We tried to get them to put them in some sort of order so it was easier for us to track for processing. But we understand that they were as busy as we were….' (Watkins transcript).

Figure 9.4: Poster that Watkins and her team created to send out to different laboratories to encourage them to send in samples in the right way. The poster outlines the journey the sample takes to the unit and how RNA can degrade if not processed in a timely manner. The team at PenGu also regularly sent out newsletters to underline the importance of appropriate storage and getting samples to it quickly to prevent degradation of the RNA (Watkins transcript).

Likewise, the team at the Public Health Laboratory in Bristol also wrestled with some difficulties in the way that samples arrived. Stephanie Hutchings and Hannah Pymont, trainee clinical scientists who volunteered to help the COG-UK effort there, remember the many different shapes and sizes of the containers that the samples came in which made it difficult to process internally. Pymont recalls 'Some sites were sending us the big primary tubes that they'd originally swab the patient with, some people were aliquoting it off into smaller ones to make it easier to send.' She says 'To start with every sample that was sent to us had to have a particular reference form attached to it with the patient's details, which was realistic when you only had a few samples coming in from each site, they could fill those in. But then there began to be piles and piles and piles of paper in category three [a biosafety laboratory], and we'd end up having to reject samples because things hadn't been filled in properly. We'd occasionally phone back to the sites, but it was just not realistic at that time, especially during the April/May peak in 2020 when things really began to be really bad, and then winter hit after that. It just got to the point that it wasn't manageable anymore.' To bring some order to the system, they created a form on an electronic requesting system which already had details in place for all of the local hospitals. With that, Hutchings says, 'the majority of our samples were okay.' She points out that there was always going to be a percentage of samples that 'have labelling errors or certain things' but this was not more than they would normally expect' (Hutchings and Pymont transcript).

Figure 9.5: Photograph of Stephanie Hutchings (back) and Hannah Pymont (in front). Credit: Pymont. By early 2021 the laboratory in Bristol had sequenced over 1200 samples (Pymont).

Figure 9.6: Poster retweeted by Hannah Pymont, 20 July 2021. The poster was created by Mandy (surname unknown), based at British Columbia's Children's & Women's Hospital and tweeted out by the Pathology Department at University Hospitals Birmingham based at Queen Elizabeth Hospital.

Many of the problems encountered were not surprising given how over-stretched diagnostic laboratories were at the start of the pandemic. Some idea of how overloaded such staff were is captured in the interview with Dr Alex Trotter who spent the first few months of the pandemic collecting samples from Norwich and Norfolk University Hospital (NNUH) for sequencing at the Quadram Institute. He described it being 'the wildest time the hospital lab has ever been in, they were completely overrun compared to normal… They borrowed staff from other departments in the hospital, bank staffing, and students just finishing, anyone they could get in to cover all the shifts they had, all of them suddenly having to process such a great increase in the amount of samples that they had.' Given this situation he says 'I was always very conscious that I was in the way of about 30 people rushing around trying to desperately run a hospital lab. It was tough' (Trotter transcript).

The Sanger Institute team also faced obstacles in terms of how samples came in. Initially the plates they received from the Lighthouse labs were 'loaded quite haphazardly into boxes' (Kwiatkowski transcript). Ewan Harrison also remembers there were times when the 'plates would arrive, but the maps of which sample was positive, which sample was negative, wouldn't arrive. So basically, you had samples sitting there that you couldn't do anything with until these maps arrived late' (Harrison and Jermy transcript). In addition to this, the data could sometimes be incomplete, which caused delays. The complications this caused are outlined by Professor Dominic Kwiatkowski. He recalls 'in the early days whilst we might know where a positive sample was, we wouldn't necessarily know what part of the country that sample had come from because there were multiple different data feeds. One was from the lab itself saying “we've tested the sample, so in this plate of samples, this one is positive”. And there was another data feed saying of all the samples in that plate, this one comes from Runcorn, and this one comes from Nottingham and so on' (Kwiatkowski transcript).

Other small issues could also confound the work. Dr Ian Johnston, the head of sequencing operations at the Sanger Institute, for example, remembers problems reading the barcodes on the plates because they were not suitable for the freezer, or the barcodes not oriented in the right place for scanning which slowed things down. At one point there was also a problem with the seals on the plates. Such problems were resolved once a member of the Sanger Institute staff became embedded in the Lighthouse lab in Milton Keynes (Johnston transcript).

Logistics and transport issues

Getting samples transported for sequencing posed a major logistics challenge. When Ludden first started working on the logistics for COG-UK she says 'it was much like a blank piece of paper. It was essentially, what testing is being done? What samples are being taken? How can we even transport those samples? What is the cheapest way to do it? What kind of boxes could be used? Do we really need dry ice?” (Ludden and Blane transcript). The dry ice was critical to preventing the RNA degrading which could severely hamper the quality of the sample and the sequence result (Watkins transcript). Just how hard the logistics were was also flagged up by Dr Sam Robson at Portsmouth University. According to him this was 'probably among the hardest things to arrange.' He points out 'It was not just the difficulties of getting samples from point A to point B, that's fairly straightforward if you've got a decent courier doing the transport for you, but dealing with everything that comes with it' (Robson and Beckett transcript).

Hutchings similarly emphasises that while logistics might seem relatively straightforward, it involved a lot of work. For example, she says, 'Something like dry ice, arranging for it to turn up, making sure that you have a driver, and making sure that it's received at the other end, actually is quite a large undertaking, but on the surface seems very simple.' When the team in Bristol first got involved with COG-UK work, the Sanger Institute set up a service so that they could order dry ice, which made life a lot easier (Hutchings and Pymont transcript).

The complexity of logistics was also highlighted by Dr Nabil Alikhan, a bioinformatician involved in the COG-UK work at the Quadram Institute. 'That's always been a problem regardless of whether a single organisation has been handling it [the samples] or whether it's been passed around multiple organisations. There is always a time lag because you have to ship things via courier from a hospital site all over the country to someplace, that's always a problem' (Alikhan transcript).

In order to encourage testing sites in hospitals to send as many of their positive samples as possible, the team at COG-UK's hub helped coordinate couriers to collect samples (Ludden and Blane transcript). For Dr Sarah Buchanan, who was contributing to the COG-UK operation from Bournemouth, this was enormously helpful. She recollects, 'A courier would turn up with their dry ice, collect our plates, and it would be shipped off immediately. So hats off to whoever did that, it was superb from the start' (Buchanan transcript).

Many of the couriers were themselves working very long hours. Hutchings recalls on one occasion calling a courier at 10pm, when she was working on-call, to arrange for the collection of a sample and being told they first needed to drive to London and then Plymouth before they could get to Bristol. She says 'It was four in the morning before they actually arrived. I think that the courier services as well really had to adapt and change their work depending on the amount of samples that we had and the amount of requests. So, it was a huge undertaking by all' (Hutchings and Pymont transcript).

In the majority of cases the samples arrived safely. But sometimes they took a circuitous route. For example, in Portsmouth, Beckett remembers at the beginning samples being sent by taxi, 'so then every day a different taxi driver would arrive and have to try and find our lab within the whole hospital complex. In the end Sharon [Glaysher] designed a really nice little map with special directions to help people find us a bit more easily, which we sent to the sites so that they could stick them to the boxes they were sending to us' (Robson and Beckett transcript). Similarly, Kwiatkowski has a vivid memory of one weekend being phoned up because there was a concern 'that a bunch of samples that should have come from the south coast had gone' and were 'sitting around in some sort of FedEx package that might get delivered to someone by mistake. Fortunately, they were not biologically active samples, so it was all okay' (Kwiatkowski transcript).

At the University of Northumbria, the team 'realised the cost of shipping samples, two, five, 10 miles down the road with dry ice was going to cost a fortune' so instead one of them drove 'around from lab to lab, picking up the samples and bringing them back for prep and sequencing.' They could do this because the university agreed for them to use the vans then not in use by the security staff. Professor Darren Smith, who led the team, remembers that the driving was done by 'Whoever was in the laboratory who didn't have a pipette in hand. Whoever was free, whoever had finished their parts of the pipeline'. He recalls, 'We had all the health and safety documented and training so we could carry dry ice and samples. There were places like Gateshead and the Royal Victoria Hospital which are around two or three miles away which is a five minutes' drive. So, if there were any problems, or a possible outbreak at one of our 7 sites or we thought a shipment was going to be delayed or it was easier for us to go and pick them up, we just did that. North Cumbria is about an hour, hour and a half's drive, and even then some of the guys wanted to do it because it's quite a scenic drive from Newcastle to North Cumbria as you go through the Lake District on the way through' (Smith transcript).

In some cases, samples were picked up in person. This could be time-consuming as is testified by Trotter who did such collections from the NNUH for the Quadram Institute. The hospital is on the same campus as the Institute just five minutes walk away. Provided with a list by the hospital, he had major problems at the beginning finding the samples at the hospital 'because everything was done ad hoc at the time…I might get the information that I've got a list of samples and I know three fridges they might be in, and I would think “wish me luck”.... So I'm going to rack in the fridge … and there were 2000 samples in there, most were negative. I spent a lot of hours rooting through their fridges…. I think I've sort of blocked out a lot of things from the first year of my memory like how difficult it was going to the hospital to collect samples all the time (Trotter transcript).

The same issues cropped up in terms of the collection of samples from the Public Health diagnostic laboratory at Addenbrooke's Hospital, which were sequenced by Ian Goodfellow's laboratory at the University of Cambridge. He remembers that the process often took two people three hours a day to find the samples stored in a freezer out of the thousands of samples the hospital received each day from the whole of the East of England and the surrounding area (Goodfellow transcript; Török transcript).

Sampling framework

One of the central discussions at the start of COG-UK was how it would achieve a 'balance between local sampling and surveillance sampling.' Some idea of the thinking that went behind this is summed up by Julian Parkhill. At the very first meeting in London he remembers there was a lot of talk 'about local reactive sampling to look at things of local importance, are you seeing hospital outbreaks, are you seeing outbreaks in particular local areas? That needs the ability to go in and sequence local samples.' This needed to be balanced against the objectives of 'the bioinformaticians, phylogeneticists' who 'wanted to be able to understand the progress of the pandemic across the whole UK, and for that to happen you need unbiased sampling. You need to be able to know that the samples you are dealing with are an unbiased collection, an unbiased sampling of all the things that are going across the UK, and just sticking together ad hoc collections that people have done from local use will give you the wrong answers. It is also important for questions like how fast is the pandemic growing, which you can get from these phylogenetic trees that you can't get in other ways' (Parkhill transcript).

To address the issues, Parkhill was invited by Peacock to develop a sampling framework that would 'kind of square that circle of making sure that there were useful things being done at a local level, but also the data would be useful at a UK-wide level.' Based on several conversations over two weeks he came up with the idea that each centre would use a specified proportion of their COG-UK funding 'towards sequencing random samples as they come through the door, which was important for national surveillance.' Sequencing sites could then use the remaining funding to address questions of local importance. The crucial thing was for each centre to make sure these were labelled so that when they were submitted to central databases the surveillance samples could be pulled out. Parkhill emphasises that 'It was the same fixed amount of effort in all of the sites going into surveillance which gave us that broad, UK-wide framework of surveillance' (Parkhill transcript).

How the framework operated in practice within Scotland is outlined by Dr Ana da Silva Filipe, who helped manage the COG-UK operation at the Centre for Virus Research in Glasgow. She remembers 'we defined what the sequencing capacity was between Edinburgh and Glasgow. Then we took into account the number of people in different health boards and tried accordingly to define what would be the number of samples in terms of percentage that we would be asking each of the diagnostic labs to send to us. That aspect was as random as it could be. There was no pre-selection of samples. Then in parallel, whenever there was a suspicion of an outbreak, we would target all the samples within that outbreak for sequencing to characterise that outbreak in terms of transmission' (da Silva Filipe transcript).

At PenGu in Wales, Watkins recalls the team had meetings to 'keep an eye on where samples should be coming from that are of interest, what outbreaks were particularly of interest that we needed to ensure the samples were sequenced.' Her recollection is 'The epidemiologist would send us a spreadsheet of outbreaks with patient-identifying information on, whether that be a hospital number or a lab number. And then we would search those to ensure that they were going through the sequencing, and roughly when the results would be available. We would then forward that on to our bioinformaticians, who would then utilise that spreadsheet with the read information, to then ensure that that sequencing data then got forwarded on to the epidemiologist. It was quite manual to start off with. But our bioinformaticians are fantastic. And by the end of it all, it was a fully automated process, which would then go to Public Health Wales, who would then disseminate information to all of the epidemiologists throughout Wales.' This provided a good check on a weekly basis 'to ensure that we were processing and getting through the system what needed to be got through the system' (Watkins transcript).

Back at the Cambridge administrative hub, the practicalities of getting samples between different centres demanded 'responding to an evolving situation.' Ludden explains at the beginning 'we were getting a lot of requests from the public health agencies referring to outbreaks and the need to get the samples from one point to another as quickly as possible. So it was quite responsive and we did continue to get that throughout. But as time went on, especially as we built the team, we tried to make it more systematic and become more organised by arranging collections from sites every week. Also over time, we began to realise that our approach wasn't sustainable because the case numbers became too high. We couldn't just keep responding to outbreaks and we needed to build in some capacity for additional high impact urgent requests. We needed to be more forward-looking than constantly look back at the old outbreaks, we needed to see what was happening now, because some of those outbreaks may have been three weeks previous and that really wasn't as useful as looking at what happened in the last week' (Ludden and Blane transcript).

Because the Sanger Institute received many thousands of test samples a week from the Lighthouse labs, it had to develop its own strategy to ensure that it got an even geographical spread across the country. This was done with the help of the informatics team who wrote a computer algorithm to programme the liquid handling robots 'to pick certain samples from all of these positives' that came to the Institute. The algorithm was invaluable because at its peak, in January 2021, the Institute was receiving over a million samples a week (Langford transcript).

Storage and destruction

Just as logistics was a challenge, so was the storage and retention of samples in case sequencing needed to be repeated, or in the event that non-sequenced samples became part of an outbreak investigation and a request was made for them to be sequenced (Marjanovic Final Report; Glaysher transcript; Watkins transcript; Temperton and Michell transcript). The Sanger Institute, which received the largest quantity of samples each day, addressed this by installing large chilled containers in its car park. Elsewhere, other centres did not have this option. In Northern Ireland, for example, Queen's University Belfast had no storage of its own and returned any unsequenced pillar 2 samples to Randox [a company involved in community testing in the UK] for storage (Miskelly transcript).

In Portsmouth, they utilised a number of minus 80 freezers already available in a dedicated research laboratory facility at the hospital which they usually used to store samples from trials. Their ability to store COVID-19 samples in these freezers was helped by the fact that at the start of the pandemic they were relatively empty because the hospital restricted the number of patients admitted and a lot of research was put on hold. But over time this situation changed. Glaysher remembers 'It went from just having a few samples to nearly a whole freezer that was completely full. You'd open the freezer and then there were thousands of samples in there. Capacity-wise we were worrying towards that first winter period, could we analyse them quickly enough before we reached maximum capacity?' Fortunately, once the samples were processed ready for sequencing, Glaysher says, 'their footprint was smaller' which made it easier to store them (Glaysher transcript).

The PenGU team in Wales was also well set up for storing the samples. Watkins explains, 'Because within our service we had been doing respiratory PCRs for many years, we had already devised a dry swab in a lysis buffer as a storage sample. So, we always had the sample to go back to, that sample would be stored. We always used to keep samples for a minimum of two years in a lysis buffer.' But even here 'That did get to be difficult with COVID with the numbers eventually.' The minimum they were able to store samples was for two months, but this still gave them capacity to go back and 'get samples that might have been associated with outbreaks or may have failed for whatever reason' (Watkins transcript).

Managing the storage of samples could pose some issues. Some of the complexities experienced in the Public Health Laboratory in Bristol are described by Hannah Pymont: 'To start with we were storing every single positive in the freezer that we could get our hands on, so anything that came up positive we were storing the original primary aliquot in our minus 80 freezers. That was because we didn't know at the time what would be needed, obviously it was early on, everyone was trying to find out more about this virus, we were sending things off for culture, various things as well. We eventually ran out of minus 80 freezer space and that was one of our biggest problems.' The situation changed once the Alpha and Delta variants became more prevalent, at which point they did not need to store 'everything positive coming through because by then' they had 'got a lot of data on the virus.' At that point, she says, the historical samples were 'still very useful to look at in terms of how it might evolve in the future, but that is just the volume of aliquot storage. Not only did that require freezer space, it required boxes to store them in. We had hours of fun just trying to get hold of cardboard because there was a cardboard shortage' (Hutchings and Pymont transcript).

Because so many samples were coming in, the Bristol team had to set up a storage system within their electronic laboratory system. A similar system was also put in place at the Sanger Institute. This made it possible to scan in each sample based on a barcode to generate information about the whereabouts of where it was stored so it could be pulled out when needed. According to Hutchings, the system in Bristol 'was better than trying to manually scan each one individually to see what result it had come up with, and less prone to human error at that point as well' (Hutchings and Pymont transcript).

In accordance with ethical requirements, the samples could only be stored for a limited time. This meant each centre needed to have storage capacity in place and also the means to destroy samples in accordance with guidelines. Destroying samples not only helped free up capacity in places with limited freezer storage, but also for the Sanger Institute which handled over 26 million samples in under two years (Brooklyn transcript). Unlike pillar 2 samples that arrived in 96 well plates and were of the community (Lighthouse) testing labs, hospital samples could not be so easily discarded and were often sent to biobanks whose purpose is to collect, store and store large amounts of human material for medical research (Ludden and Blane transcript).

Sequencing technology

Each COG-UK centre had the freedom to choose which sequencing platform to use for their samples. This was largely determined by what equipment they already had in place, which they subsequently supplemented with other machines funded by COG-UK and follow-on funding from Test and Trace. It also depended on the volume of samples to be sequenced and speed they had to be turned around. One of the advantages of the ARTIC protocol was that it could be adapted to many different sequencing platforms (Darby transcript).

Centre/sequencer modelTotal

Belfast Health & Social Care Trust/Queen’s University Belfast1

MiSeq1

Northumbria University9

GridION1

MinION4

MiSeq2

NextSeq1

PacBio Sequel I1

Public Health England Colindale2

Illumina (unspecified)1

GridION1

Public Health Wales & Cardiff University5

GridION1

MinION2

MiSeq1

NextSeq1

Quadram Institute7

MinION1

MinION4

NextSeq1

PromethION1

University College London10

GridION1

MinION4

MiSeq2

NextSeq 500/5502

PromethION1

University of Birmingham5

GridION1

HiSeq1

MiSeq1

NextSeq1

PromethION1

University of Cambridge2

GridION1

MinION1

University of Edinburgh5

GridION1

MiSeq1

MinION3

University of Exeter5

GridION1

MinION4

University of Glasgow Centre for Virus Research7

MinION3

MiSeq2

NextSeq2

University of Liverpool17

Iseq1

MinION10

MiSeq4

NovaSeq1

PacBio Sequel I1

University of Nottingham3

GridION1

MiSeq1

PromethION1

University of Oxford5

MinION1

MiSeq2

NextSeq1

NovaSeq1

University of Portsmouth3

GridION1

MinION2

University of Sheffield5

GridION1

MinION4

Wellcome Sanger Institute43

GridION1

MinION9

MiSeq6

NextSeq1

NovaSeq19

PacBio Sequel I1

PacBio Sequel II5

PromethION1

Manufacturer/machineTotal

Illumina57

HiSeq1

Iseq1

MiSeq23

NextSeq8

NextSeq 500/5502

NovaSeq21

Unspecified1

Oxford Nanopore Technologies69

GridION12

MinION52

PromethION5

PacBio8

PacBio Sequel I3

PacBio Sequel II5

Figure 9.7: Total number of sequencers at each COG-UK centre and by manufacturer. CLICK ON a centre or manufacturer for details of specific sequencer models. Credit: Annex 6, Marjanovic Indexes. In many cases, different centres used a mixture of sequencers. MinION, GridION and PromethION are nanopore sequencers, NextSeq, MiSeq, NovSeq are high-throughput sequencers produced by Illumina and PacBioSequels are high-throughput machines made by PacBio.

The Sanger Institute had a bank of high-throughput Illumina sequencers and worked with the company to push their sequencing capacity to the maximum. Together, they halved the sequencing run time down to less than 24 hours (Johnston transcript). But Illumina machines were not suitable everywhere. One of the downsides is that there needs to be enough samples for it to be worthwhile to put a run on using a high capacity sequencing instrument. Some places, like the Centre for Virus Research in Glasgow, used Nanopore sequencing which catered to lower volumes and was more practical for investigating outbreaks, and high-throughput Illumina machines when they needed to sequence large volumes (Thomson transcript; da Silva Filipe transcript).

Sequencing capacity

As the pandemic progressed, so did the COG-UK sequencing capacity. Dr Matthew Bashton, a bioinformatician at Northumbria University, argues this was largely down to innovations made by different sites to improve the throughput on their sequencers. It was also helped by the development of the LoCost Protocol by Joshua Quick and his colleagues. For example, at the start Bashton remembers only being able to put 24 samples on a flow cell used in an individual Nanopore sequencer, which rose to more than 50 samples. Similarly, the capacity on the Illumina machine was increased. Bashton points out that sequencers were not the only factor driving capacity. What was also needed were 'more PCR machines' and lab workers to work those PCR machines, and then you need more consumables, more plastics, more pipette tips. In summary, he says, 'just having a bigger sequencer won't solve the problem on its own, there's stuff that supports that as well. It was a gradual increase, which made it easier than having to immediately hit that capacity' (Bashton transcript).

The turnaround time from sample collection to upload of data also improved across consortium sites. Overall, the average turnaround time went down by 70%, from 20 days, recorded in April 2020, to just 6 days in June 2021. The turnaround time was measured from the arrival of an extract from a positive PCR test to the first interpretation of genome data. This was accompanied by a reduction in cost which went down by approximately 30%, from £56 to £40 per sample. This sum excludes labour and overhead costs (Peacock 2021; Marjanovic Final Report).

References

Jaquith, K (28 Jan 2014) 'The history of the microplate'. Back

Marjanovic, S, Romanelli, R, Claire-Ali, et al (2022) Evaluation of the COVID-19 Genomics UK (COG-UK) Consortium, Final Report, RAND Europe.Back

Marjanovic, S, Romanelli, R, Claire-Ali, et al (2022)Evaluation of the COVID-19 Genomics UK (COG-UK) Consortium, Annexes, RAND Europe.Back

Mishra, P (7 Jan 2021) 'The history of microplates', Medical Life Sciences News.Back

Peacock, S (22 Jan 2021) ‘Reflections on the achievements of COG-UK’.Back

Pymont, H (Oct 2021) 'COVID-19 sequencing… or how I learned to stop worrying (sort of) and enjoy microbial genetics', Trainee News, 673.Back

Interview transcripts

Note: The position listed by the people below is the one that they held when interviewed and may have subsequently changed.

Interview with Dr Nabil-Fareed Alikhan, Bioinformatics Scientific Programmer at the Quadram Institute.Back

Interview with Dr Matthew Bashton, Computational Biologist, Northumbria University.Back

Interview with Mrs Tanya Brooklyn, Genomics Surveillance Implementation Manager, Wellcome Sanger Institute.Back

Interview with Dr Sarah Buchanan, Lecturer in immunology, Bournemouth University.Back

"Interview with Professor Alastair Darby, Principal Investigator and Co-Director of the Centre for Genome Research, University of Liverpool (interviewed 10 Jan 2023, unpublished transcript).Back

Interview with Dr Ana Da Silva Filipe, Research fellow, NGS Facility Manager, Centre for Virus Research, University of Glasgow.Back

Interview with Dr Sharon Glaysher, Specialist Biomedical Scientist who manages Portsmouth Hospitals University NHS Trust's Research Laboratory.Back

Interview with Ian Goodfellow, professor of virology, University of Cambridge (interviewed 15 Dec 2022, unpublished).Back

Interview with Stephanie Hutchings (Trainee Clinical Scientist in Infection Sciences), Hannah Pymont (Qualified Clinical Scientist, Trainee Consultant Clinical Scientist in Microbiology), Public Health Laboratory, Bristol.Back

Interview with Dr Ewan Harrison (Deputy Director COG-UK and UKRI Innovation Fellow, Wellcome Sanger Institute, Senior Research Associate, Department of Medicine, University of Cambridge) and Dr Andrew Jermy (External Communications Advisor COG-UK).Back

Interview with Dr Ian Johnston, Head Of Sequencing Operations & R&D, Wellcome Sanger Institute.Back

Interview with Professor Dominic Kwiatkowski, Head of Parasites and Microbes Programme at the Wellcome Sanger Institute in Cambridge and Professor of Genomics at University of Oxford.Back

Interview with Dr Cordelia Langford, Director of Scientific Operations, Wellcome Sanger Institute.Back

Interview with Dr Catherine Ludden, Director of Operations, COG-UK and Beth Blane, Logistics Manager for COG-UK, Research Assistant in the Department of Medicine, University of Cambridge.Back

Interview with Dr Julia Miskelly, Manager of the Genomics Core Technology Unit, Queen's University Belfast.Back

Interview with Professor Julian Parkhill, Department of Veterinary Medicine, University of Cambridge.Back

Interview with Sharon Peacock, Professor of Public Health and Microbiology in the Department of Medicine, Cambridge University and Executive Director of the COVID-19 Genomics UK (COG-UK) Consortium.Back

Interview with Dr Sam Robson, Principal Research Fellow (Bioinformatics), Angela Beckett, Specialist Technician (Research), Faculty of Science & Health, School of Biological Sciences, Centre for Enzyme Innovation, University of Portsmouth.Back

IInterview with Darren Smith, Professor of Bacteriophage Biology, Northumbria UniversityBack

Interview with Ben Temperton, Associate Professor of Microbiology, University of Exeter; Dr Steve Michell, Senior Lecturer in Molecular Microbiology, University of Exeter and COG-UK Principal Investigator.Back

Interview with Emma Thomson, Professor in Infectious Diseases, Centre for Virus Research, Glasgow University (interviewed 11 Jan 2022, unpublished).Back

Interview with Dr Estée Török, Consultant at Cambridge University Hospitals NHS Foundation Trust and Senior Visiting Fellow University of Cambridge.Back

Interview with Dr Alex Trotter, Bioscience researcher, Quadram Institute.Back

Interview with Joanne Watkins, Senior Biomedical Scientist, Deputy Head, Pathogen Genomics Unit, Public Health Wales.Back

Interview with Peijin Zhang, Manager of the University of Sheffield's COG-UK technical team .Back

Respond to or comment on this page on our feeds on Facebook, Instagram, Mastodon or Twitter.