COG-UK’s hub gets moving

From the start it was clear that COG-UK's distributed sequencing network would not function without a central administrative hub to help oversee what was happening, organise meetings and bring people together. Peacock quickly stepped up to take on this task. Looking back, Peacock realises she 'could have looked for someone else to run COG, but actually I felt that I had to lead it at the time.' As she argues, 'There had to be someone to run the administrative hub for the spokes'. Based on this, in April 2020, she decided to hand over her role as director at the National Infection Service so that she could focus all her attention on setting up COG-UK (Peacock transcript).

Early COG-UK management team

Having no formal COG-UK team at the start, Peacock initially turned to members of her research team for help. One of the advantages she had in doing this was that many of them had already been working with her for at least a decade so she was able to say to them 'Can you just put everything down and help me develop COG-UK'. She recalls, 'Every one of them thought about it quite hard, but they had enough trust in me to do it. They all said yes… Everybody absolutely wanted to make that contribution.' According to Peacock 'it was a defining moment in my group, because they had to have faith and trust in me to follow me… When I asked my group to follow me, I didn't have a huge budget.' It also involved them having to leave their research programmes, although, as Peacock points out, they became limited in what they could do because labs had closed due to lockdown (Peacock transcript).

One of the first to become heavily involved in helping Peacock with setting up COG-UK was Dr Ewan Harrison, who was working on genomics to track methicillin-resistant Staphylococcus aureus transmission and understand drug resistance with Peacock. Harrison rapidly became central to co-ordinating the development of COG-UK because, as he says, Peacock was then in the deep end of the COVID response and 'pretty much everything at that point.' One of the first things he did was join a meeting at the Sanger that afternoon where they mapped out on a whiteboard the 'whole process for sequencing from end to end' and 'how the samples would come in.' He then 'basically rang up everybody who had been in the meeting at Wellcome, and had about half an hour to an hour on the phone with them over a period of about a week to ask “What do you need? What's the overall plan? What do you think you're going to be able to do?' Harrison remembers, 'From those phone calls we very quickly got sent documents from people, and then we started to bring in other people'. A lot of this concerned computational issues and thinking through the sequencing pipeline which he did in coordination with Professor Tom Connor at Cardiff University and Professor Andrew Rambaut at Edinburgh University (Harrison and Jermy transcript).

Figure 5.1: Photograph of Dr Ewan Harrison, head of Respiratory Virus and Microbiome Initiative. Credit: Wellcome Sanger Institute. Fascinated by microbes from a young age, Harrison became interested in infectious diseases after reading Paul de Krieff's book 'Microbe Hunters'. Harrison completed a doctorate at Leicester University. His doctoral research focused on the genetic mechanisms unpinning the evolution and pathogenesis of Pseudomonas aeruginosa. He then became a postdoctoral fellow at the University of Cambridge, focusing on the use of genomics to track the transmission of MRSA between humans and animals and between hospitals and the community. In 2018 Harrison joined the Sanger Institute as part of the Health Data Research UK where he mounted a study to investigate the biological basis of the persistent colonisation of S. aureus in the nose. This involved collecting samples from large cohorts of blood donors using microbiology, electronic health data, genetics and microbiome studies (Harrison profile).

In addition to Harrison, Peacock was also assisted by Professor Gordon Dougan, an expert in vaccinology, infection and immunology. He proved invaluable, giving support where needed to getting COG-UK off the ground. Much of their conversation, conducted over Zoom, revolved around working out what each partner was doing and who were their key contacts not only scientifically but also in terms of administration and the university.

At this point the funding proposal was still being pulled together. Harrison remembers that at that moment 'some bits were a bit more thought out' than others. One of the most thought out plans he argues came from Connor, who had the advantage that some years before he had been involved in building a pathogen genomic service in Wales. Connor also sketched out a plan for 'how the computational stuff would work in CLIMB'. What also helped was the ideas that 'came out of the initial brainstorming meeting at Wellcome'. This still left some things to be developed 'on the hoof'. In the end, these different components had to be turned into 'a funding proposal that was linear and made sense' (Harrison and Jermy transcript).

Financing the work

An enormous part of the early work involved putting together a budget. This was not easy. A lot of thought went into how to divide up funding so that it was fair and everyone got what they needed to achieve. This was especially important to Peacock. For her this matter was not only about making sure people got what they needed but also about making sure the money was used properly because it was public money. It came from UKRI [UK Research and Innovation, a government funded body] and the Wellcome Sanger Institute (Peacock transcript). In order to work out how to divide up the budget, Harrison asked everyone how much it would cost them to sequence a sample. He recalls that at first, 'we got crazy numbers' but he subsequently got more realistic numbers (Harrison and Jermy transcript).

Another person who helped sort out the finance was Professor Patrick Maxwell, the head of the School of Clinical at Cambridge. Approached by Peacock to help set up the administration of COG-UK under the auspices of the University of Cambridge, Maxwell remembers that the COG-UK grant came with 'some pretty major responsibilities and expectations'. One of the most significant pressures was that 'from the beginning the Government's point of view was that they were not prepared to pay normal overheads on it. In other words, they were not planning to fund the full costs of the research, or what the Government-agreed university accounting system shows are the full costs of research; they wanted it done for less than that'. Adding to this there were other stresses. One of these, for example, was the 'VAT liability between different parties' on the grant.' At the time Maxwell thought they 'could just take a sensible overview of that' but basically the institutions landed up taking 'substantial risks' in that area (Maxwell transcript).

Most of the funding secured was earmarked to cover the cost of sequencing at the different sites. This meant that there was little left over for paying costs of salaries or equipment. Almost all of the people who participated in the consortium did so voluntarily with no payment to themselves. What drove them to contribute is highlighted in the words of one person transcripted for the RAND Report who commented, 'I think it's basically people dedicated to wanting to help. It was never about money – nobody made any money from this' (Marjanovic). The high degree of altruism is also underlined by Parkhill. As he says, 'Almost all of the people who were actually running COG-UK, none of us got any funding for ourselves' (Parkhill transcript). Dr Joshua Quick also indicates that COG-UK was 'entirely driven by goodwill, because we didn't get anything' (Quick transcript). Just how remarkable the attitude was is reflected in the fact that Parkhill remembers 'speaking to an American journalist about this, who expressed a huge surprise that COG was built on volunteer efforts. [laughs]. Nobody was paid for doing it. And I thought, well that's just the way you do things in the UK [laughs]' (Parkhill transcript).

Many who contributed to the COG-UK effort were 'only able to do so thanks to the support provided at risk by their institutions'. This included funding the 'electricity requirements of the buildings and machines used for COG-UK work' (Marjanovic). In many cases universities were willing to fund sequencing work before any money came in from COG-UK. Dr Sam Robson, for example, from the University of Portsmouth, remembers that Professor Sherria Hoskins, the dean of the Faculty of Science and Health, 'was very supportive of anything that we could do. She provided us with a bit of funding which was matched as well by the Centre for Enzyme Innovation, which is the research centre [at the university] that I'm a part of. With this money in hand we were able to get things started and start buying bits and pieces of equipment' (Robson and Beckett transcript). How supportive institutions were is also reflected in the transcript with Dr Thushan de Silva at the University of Sheffield. His departmental manager set up an account for him 'to use on trust that COG-UK would happen.' He points out 'I think at that time we hadn't even seen a contract or anything like that. She set up an account I could use to just help with the pandemic efforts. I think we probably used about £50-60,000 on that account on trust before the COG-UK money and the contracts kicked in and we could reimburse the department (de Silva transcript).

From the outset, speed was of the essence to get COG-UK up and running. This inevitably meant many institutions and people decided to take the risk of getting involved before any contractual arrangements had been put in place. In many ways this meant people had to have faith that they would subsequently get the money. Some of the delay was down to the fact that at the start COG-UK was not a legal entity and also did not have sufficient 'administrative, operational and logistics support staff' to implement contracts (Marjanovic).

A lot of the researchers who participated in COG-UK did so on the back of salaries they received from their universities and research grants. Quick recalls, 'Most people were seconded from their posts, which meant they were still getting paid from their normal jobs. But they weren't doing that job. I'm a UKRI fellow which was paying my salary, but I was doing COG-UK work. The same can be said of all the other PI's in COG-UK who also didn't get money. It was only the lab staff that got paid by COG-UK. So basically, it would have been impossible if it wasn't for the fact that our funders were just turning a blind eye to the fact that we were working on something else.' The response of the funders underlines how much of an emergency everyone found themselves in. Quick explains, 'You're relying on the sort of wartime mentality where anything goes. I wouldn't now just be able to go off for 18 months, but at the time I just emailed my funders at UKRI and said that in the terms of my contract I'm allowed to work six hours a week on some non-fellowship activities, but I'm basically working full time on coronavirus sequencing. And they were just like, "Yeah, cool. We're happy with that. Good luck". Everyone in some ways felt lucky that they could actually contribute, because there were some people who were not able to do anything. For example, you could be an experimentalist getting paid and sitting at home because you couldn't do any experiments. You were just doing nothing. So it was better than that' (Quick transcript).

One of the issues with the budget was that it only covered the 'core number of people who'd been involved in that Wellcome meeting who had agreed to be involved'. But following the submission of the proposal a number of other places joined, which meant that their costs were not included in the budget. What was amazing to Harrison was some of those that came on board later 'basically came in at risk, because they weren't named on the funding application' (Harrison and Jermy transcript). This included Professor Eleni Nastouli, a virologist based at University College London Hospital, whose group did not receive funding from COG-UK. Their sequencing was carried out using funding they had from the NHS and other grants. One of the reasons they agreed to participate was because they 'wanted to be part of the national effort and contribute to that' (Nastouli transcript).

The same was true for Dr Beatrix Kele and Dr Maria Teresa Cutino, clinical scientists based at Barts Health NHS Trust, who used money from their existing budget to get the sequencing going. At the end of the day Cutino says they 'just thought 'let's just get what we need to start doing it and then we'll apply for money from the [Barts] charity'. For them, one of the important factors driving them forward was also scientific curiosity and recognition that participation could help them expand the hospital's genomic sequencing capacity, which at that point was minimal (Kele and Cutino transcript). A similar view was taken at the Royal Devon and Exeter NHS Foundation Trust where Dr Ray Sheridan, a consultant in geriatrics, had already been thinking about how to get the hospital more involved with pathogen genomics before COVID-19 hit (Temperton and Michell transcript).

Pulling everyone together

Figure 5.2 This cartoon 'The Fates' drawn by Alex Cagan captures the spirit of the way everyone pulled together within COG-UK.

In addition to sorting out the budget, a lot of the early work at the hub involved working out who would be responsible for what. This was not straightforward because the consortium included a lot of eminent scientists each of whom had their own opinions and priorities. What helped hold them together was Peacock's strong diplomacy and also Dougan working in the background to get everybody on board. Dougan had attended the initial Wellcome meeting and then chaired the Governance Committee that had external members and was the person who drove the documentation and agenda. The aim of the Governance committee was to make sure COG-UK did everything correctly and made good use of public funding (Peacock transcript).

Much of the organisation of COG-UK was driven out of the University of Cambridge because Peacock held a position there and that was where the grant was held. Getting Cambridge to come on board was not a foregone conclusion. As Maxwell points out, 'Holding these very large multi-institutional grants is quite a lot of work for the institution… It was literally hundreds of person hours for lawyers and contract negotiators. I had to assure the institution that it was the right thing to do. And they were not always sure it was.' Part of the issue was the fact that nothing like COG-UK had been done before, so its grant was not like a standard grant and Cambridge did not have all the necessary infrastructure in place to handle it (Maxwell transcript).

An added pressure was the speed with which everything happened and the drive to get sequencing off the ground as fast as possible before the agreements were finally signed. As Maxwell says the need to begin without signed agreements was 'clearly a very uncomfortable position to be in at times for the people involved.' He recalls, 'the funders and government agencies were putting a lot of pressure on and couldn't see why the agreements hadn't just been signed.' This inevitably put pressure on Sharon and the COG-UK team, who Maxwell explains understandably then put pressure 'on the university research office and the people actually working on the contracts.' Not surprisingly, at times this meant things could get 'a little bit edgy and quite difficult'. One of the roles Maxwell played in the process was to try and keep people calm on both sides. To Peacock's credit, Maxwell remembers that 'she was extremely diplomatic about it at all times' (Maxwell transcript).

Developing an administrative structure for COG-UK

Having secured funding to develop the Consortium, it became possible to employ a small team of people in Cambridge who could develop all aspects of the administration that was needed to support a complex and multi-faceted collaboration, together with capabilities around comms, website, and events. According to Peacock, the appointment of Dr Katerina Galai into the role of Associate Director 'started the essential process of building more solid and sustainable foundations of the consortium administrative hub'. Peacock reflects that 'Katerina was instrumental to the delivery of the legal agreements that bound the consortium together; with the development of governance and administrative structures that underpinned our success; and in recruiting a small but highly effective team that supported her and the consortium in a wide range of on-going functions' (Peacock transcript).

Figure 5.3: Photograph of Dr Katerina Galai. Credit: COG-UK. Galai completed a doctorate at Sussex University in international law, policy research and had spent the previous decade working in project management. Coming on board in August 2020, COG-UK provided Galai with an opportunity for a career change and new forms of engagement. Working through Zoom, Galai helped to put in place a team of project managers, communications experts and administrators. All of them, she says were 'instrumental to the effective functioning of the consortium and none of whom were deterred from applying for the role by their location, as the need to commute or relocate is no longer a consideration for virtual teams.' Reflecting back, she recalls the 'fast-paced nature of working in COG-UK' meant that 'the comfortable balance of planning and action has shifted in favour of pragmatic decision-making and improvisation' (Galai).

Broad partnership from the bottom up

At the outset, both Peacock and Maxwell were determined that COG-UK 'should be a truly broad partnership, without Cambridge saying, “This is ours. We need to control it.”' As Maxwell argues, this was 'important, because if you're going to get other institutions to come in enthusiastically, they need to feel from the start that they're truly involved in every aspect, and that they really will get credit for it. If it's led in a selfish way from an organisation, it's a worry from the beginning that others will be just contributing to somewhere else's glory. So we were pretty careful about that. I think many people might not realise even now that COG-UK is rooted in Cambridge. On my part at least, that was a deliberate decision' (Maxwell transcript).

Having many different partners was equitable but posed significant challenges. One of the difficulties was each partner had their own approval systems and processes. Multiple partners also presented a potential challenge when it came to decision-making. To help the process, Peacock set up a strategic oversight Steering Group, which from early on became 'responsible for making decisions.' In Maxwell's opinion, 'Sharon used it extremely effectively as a way of managing stakeholders and relationships to limit tensions between different partners. I don't think there were that many decisions to take. Often, we were presented with an update or a proposal, and it would clearly be approved. I don't think it was wrestling with tough decisions' (Maxwell transcript).

Figure 5.4: COG-UK governance structure. Source: COG-UK.

What is important to note about COG-UK was that it was intended from the start to be a network. As Peacock emphasises 'it was down to individual partners to build their capacity. It wasn't top down, it was very much bottom up'. A key component of the COG-UK culture Peacock points out was that 'we were trying to be very inclusive all the time, and tread lightly' (Peacock transcript). The degree to which this was achieved is captured in the transcript with Robson, who ran the sequencing operation at Portsmouth University. He indicated that COG-UK was 'one of the best collaborations I think I've ever been part of in the sense that there really was no feeling that there was a hierarchy to anything… I was never made to feel that I was lower down the pecking order than anybody else. I think this is very much something that can be laid on Sharon Peacock's shoulders. It was very much set up as everybody on the same level, bringing us all together for a common goal, and I think that shone through it all the meetings. I never felt that I couldn't speak up and say something.' As he says, 'If we had comments to make, they were always taken on board and I think the structure of COG-UK was inclusive as well.' He stressed that 'everything was shared around. There was no sense that there were decisions being made behind closed doors that were made outside of our control, or that people were making decisions for us, or anything of that nature. Everything was incredibly transparent, and I always felt like our thoughts and opinions on matters were valued and would be treated as such' (Robson and Beckett transcript).

Working groups

One of the key ingredients to making COG-UK function was the fact that early on Peacock set up a number of working groups which were assigned to support different parts of the overall process. Each group ran regular online meetings (Quick transcript). The aim of the groups was to aid research and support the operational work undertaken in the consortium. Initially eight operational groups were created. These covered Modelling, Phylogenetics and Display, Sample Logistics, Metadata and Patient Linkage/Epidemiology/Health Informatics, Sequencing, Data/Bioinformatics, Clinical Virology, Mutations Research, and Wastewater (Marjanovic).

For many taking part in the working group meetings there was an overwhelmingly positive atmosphere. Anthony Underwood, who represented the Centre for Genome Pathogen Surveillance in Oxford at the group looking at the development of data software, gives a sense of the spirit of what happened. He found the meetings 'Brilliant. I absolutely loved those. The interactions at that level, I suppose with the people at the coalface, doing that sort of software development, those are some of the highlights. There was such a “can do” sort of attitude, with the teams. There was a real sense of “we're in this together, trying to understand each other, trying to understand where we're coming from”' (Underwood transcript).

Quick similarly praised the sequencing working group he was asked to chair from 3rd April 2020. The group was intended to oversee 'all the protocols that were in use throughout the network' and deal 'with any specific issues relating to sequencing failures, sample failures and stuff like that.' Most of its participants were lab managers and sequencing technicians or postdocs. The discussion usually centred on problems different people were confronting and potential ways to solve them. Quick remembers, 'A lot of the time people were discussing things not directly related to the protocols. We were more thinking about doing some automation because we had staffing shortages or things like tip shortages. People would discuss ways to save tips or to implement robotic automation or to improve data quality. There was an endless list of things that we looked at. It was good for the lab people to talk to other lab people about lab things' (Quick transcript).

One of the advantages of the sequencing group was that it was 'entirely platform and protocol agnostic'. According to Quick, 'you could be using Illumina, Nanopore, both or whatever, and you could be doing any protocol you want. That allowed people to share information.' Ideally the meetings would have one or two people from every site in COG-UK. Having as many representatives as possible was helped by the fact that the meetings were held on Zoom. Frequently, the meetings had one or two people calling in from the 14-15 sites around the country. Initially the meetings were held every week but as things stabilised this went down to two and then three weeks and finally once a month (Quick transcript).

For Dr Ana Da Silva Filipe who attended from the Centre for Virus Research in Glasgow, the sequencing meetings were extremely helpful. As she says,'The fact that every week, or every two weeks, we would be in a meeting with a large number of other groups in the UK, that could potentially be facing the same problems that I was, or that someone had detected a new problem, or someone that was struggling with something and someone else in that large group would have experienced something similar, was so positive.' She believes that 'those meetings were one of the reasons for the success of COG-UK as a whole. They were very practical, very informal' (da Silva Filipe transcript).

Describing what happened at the meetings, Da Silva Filipe says, 'Normally, there was an agenda. There were some pre-discussions before the Zoom meeting and those would help to shape the agenda. And at the end of the meetings anyone that had additional problems could also bring them up, as we would go through each representative for each one of the sequencing groups, so they would have a chance to talk and say, “this week, this is what we are going through”. So everyone would have a chance to talk and bring up the problems and that would sometimes create a new point for discussion for the next meeting for example' (da Silva Filipe transcript).

Alongside discussing issues, the meetings provided an avenue for the management team to feed back decisions. Underwood points out that the group did not 'always agree with some of the decisions, but that wasn't necessarily our call. Our call was to, with our expertise, speed our views upwards, and then someone who is in the position to make the harder judgement call and then we have to implement that.' But he recalls 'even if sometimes we disagreed, we never felt demoralised by that I suppose because we knew that wasn't really our call' (Underwood transcript).

Figure 5.5: Winged Hermes, cartoon by Alex Cagan.

Network support mechanisms

An enormous amount of COG-UK's work was also conducted through Slack and Microsoft Teams, direct messaging systems designed to bring people together to work as one unified team. For many this proved invaluable because it provided a way to talk in real-time about what was happening and ask for advice. Dr Sharon Glasher, working at Portsmouth Hospital, points out that having this mechanism meant that 'you felt like you weren't alone. Even though we were working in our little area, within our little group, you always felt that there were people out there to support you' (Glaysher transcript). The same was true for Dr Matthew Bashton at Northumbria University. He thinks 'the existence of COG and the fact that we had a group of people, predominantly other bioinformaticians and genomicists available on Slack and things like this, actually was a massive distraction [from the pandemic]'. He says 'I think that kept me sane, it really helped' (Bashton transcript). It also really helped on a practical level. Peijun Zhang who was a laboratory manager at Sheffield University remembers, 'if I was on the evening shift and I was on my own, maybe I was doing the final sequencing part and maybe I noticed something a bit strange in terms of the concentration of the samples and I needed a second opinion, I would just post on Slack. And anyone who knows about the work or had any good advice, they could just tell me straight away' (Zhang transcript).

Bioinformaticians in the consortium frequently used Slack 'if there was an update on a code or if there was a bug in the system.' They would put the information on 'one of the Slack channels so everybody could use it' (Ludden and Blane transcript). According to Robson this Slack channel 'was very heavily used right the way through, where if you had an issue you could ask directly'. He points out that if they downloaded code 'off' of GitHub or something, you would have to send an email to the creator, and then you would only hear back when they had a chance to reply back to you. Whereas here, it's a lot more like direct messaging. Most members of COG were very active on there, so you could expect a reply back pretty quickly. So anytime there was any issue on the processing side of things, that was quite easy to resolve with a lot of help from people with a lot of experience' (Robson and Beckett transcript).

Slack was also used extensively by the management team. Dr Catherine Ludden who was in charge of operations at COG-UK often used it to distribute samples that needed to be urgently sequenced. This was particularly helpful within the context of outbreaks. In this situation she says 'we were told we've got 50 samples from this hospital that need to be sequenced straight away from this area and we would contact the sequencing sites and say exactly that. It might be “I've got 50 RNAs, or 50 primary samples, who has capacity to take them tomorrow?” Where it also proved useful was when there were shortages of consumables, which happened at various points during the pandemic. Beth Blane, for example, who was also working in operations, remembers posting messages on Slack saying 'such and such a site needs 10 micropipette tips, has anyone got any to spare? Or gloves? And another site would pipe up, “yes, we've got extra”, and they arranged to send them between them'. One of the advantages of using Slack was that it was easier to send out a message than to write an email to a group and then coordinate the response (Ludden and Blane transcript).

Figure 5.6: Photograph of Dr Catherine Ludden. Credit: Catherine Ludden, Linkedin. Growing up on a farm in Galway, the Republic of Ireland, Ludden developed an interest in living things early in her life. One of five children, Ludden was encouraged to pursue a career through the example of her mother who at the age of 42, decided to take a degree, which she did by attending classes in the evenings. Ludden studied Medical Science at Galway-Mayo Institute of Technology during which time she developed a strong research interest in antibiotic resistance. She was introduced to the world of genomics during her doctorate at the National University of Ireland Galway. In 2016, Ludden was based in the Peacock group at the University of Cambridge when she was awarded a Wellcome Trust postdoctoral research fellowship to investigate the source and transmission of AMR pathogens. This involved collecting and sequencing samples from livestock, wastewater and hospital patients and working with a wide variety of people. Just before the emergence of COVID-19, Ludden spent six-months in Sweden looking into international outbreaks of antibiotic-resistant bacteria for the European Centre for Disease Control and Protection (Stubbs).

Figure 5.7: Photograph of Beth Blane, OBE. Credit: COG-UK. Science and maths were Blane's favourite subjects at school. Completing her A levels at college after leaving school, Blane went on to have children in her early 20s and worked for 9 years in retail. Realising that she really had a passion for science, Blane decided to take a degree in biomedical science at the Open University which she did part-time over 6 years. After qualifying as a clinical biomedical scientist, Blane worked for many years at Addenbrooke's Hospital in Cambridge. In 2012 she moved over to Peacock's Laboratory to become a research assistant to support the AMR work. For Blane, the pandemic completely changed the nature of her work. Normally assisting with studies on detection and sequencing of AMR bacteria, which involves a lot of hands-on laboratory work which she particularly enjoys doing, Blane switched to helping COG-UK navigate the complexity of the diagnostics network in the UK and coordinate the flow of samples to different sequencing sites (COG-UK May 2021; COG-UK Jan 2023).

But not everyone was familiar with Slack. Those who used it most were bioinformaticians who were already using it before the pandemic hit. Because not everyone used Slack, the management team often duplicated their messages on email 'if something was urgent and really important that everyone needed to hear about'. They also used it when there was a site they knew that was not looking at Slack, and on occasion they also used the phone when they were unable to get a response (Ludden and Blane transcript).

Consortium Agreement

COG-UK had multiple layers of governance which was embedded in its Consortium Agreement. Getting this agreement in place took six months of negotiation between different partners. Much of the discussion involved establishing where the money was going in terms of institutions and PIs and also sorting out who owned the intellectual property. What added to the pressure was the grant was very complex and needed to be sorted very quickly. Peacock explains, 'We had to wade through treacle to get the legal contracts completed, because getting 21 institutions to all sign up to a single consortium agreement with a data sharing agreement and a publication policy was a challenge' (Peacock transcript).

One of the biggest challenges in drawing up the agreement was how to share genomic data across the UK. As Harrison states, this was because 'There had never been an agreement between all four UK public health agencies to share pathogen data. And not only to share it with each other, to then share it with the academic research community.' A lot of the wrangling, he says, was around that (Harrison and Jermy transcript).

The Consortium Agreement needed to be able to accommodate differences across the UK. Harrison remembers, 'For example, in England, I had to write a study protocol that went via what was then Public Health England ethics. It was a study protocol that outlines everything about what we were doing, how we handle the data, what its purposes are, what we could hold in CLIMB. That covered it for England. Scotland, Wales and Northern Ireland didn't require it, they had their own internal processes that were covered. Most of that wasn't in place to begin with, but we got it all going.' But it was not all plain sailing from there. Harrison indicates that 'some labs, hospitals and NHS trusts were absolutely fine for people to crack on, but others were pedantic to the point that I probably had about ten independent calls with them, going through it time after time after time. Some of these, even after we had the ethics approval, just wouldn't believe that we were allowed to do this. It was really painful' (Harrison and Jermy transcript).

Another issue Harrison flagged up was that they 'had to work out secure ways to transfer specimen IDs from labs to the Public Health Agencies. This had to be done securely because under GDPR rules specimen IDs are patient identifiable information - this was an important problem to solve because if you don't have the specimen ID you can't link a genome to a person. Working out how to do all these things at once was very challenging'. As a result he 'had what seemed like endless hours of meetings' which he did together with Michael Chapman, Nick Loman, Tom Connor, people from public health agencies and lots of other people to work through how they could build what would work like a laboratory management system. This was highly complex. In some cases where hospital trusts had samples coming in, they regarded the information they contained as private so nobody else could view it. And yet the public health agencies needed to be able to view it to do their analysis. From his perspective the 'problem of linking a sample to a patient' was 'the hardest thing we actually did. Getting that right was really hard' (Harrison and Jermy transcript).

For Harrison, the fact that the consortium managed to get an agreement around data sharing was 'absolutely critical' to its success. He had previously been involved in 'helping other countries set up pathogen genomics' which didn't have it. One of the biggest problems there had been 'they just couldn't get anybody to agree, because people in one hospital regard their patients' data as their private fiefdom' (Harrison and Jermy transcript).

Organisations were not only being asked to take a risk in terms of data sharing. Moving samples between different sites also posed questions about the transfer of samples between different sites. Just how difficult this could be was expressed by Dr Nabil-Fareed Alikhan, a bioinformatician who played a pivotal role in helping to sequence samples for COG-UK at the Quadram Institute. Part of the problem was the fact that 'there was no infrastructure to talk about it' so it 'was a learning curve'. As he says normally 'Material transfer agreements take months to sort out between universities' but COG-UK needed to be able to 'sequence the samples in a few days for this to have any point'. From his perspective the way COG-UK managed to sort out the issue sets a major precedent and is one of its lasting legacies (Alikhan transcript).

In addition to addressing data sharing and material transfer, the Consortium Agreement also provided guidelines for publications. Reflecting the culture of inclusivity within the consortium, the agreement specified that anyone who had 'worked in any capacity to produce COG-UK's data can be added to the authorship list for COG-UK outputs.' This was also in part to help 'compensate people for their time away from other research' (Marjanovic). For academics whose career advancement relies on their list of publications this was an important incentive. Overseen by the Steering Committee the publication policy proved a bit of a challenge to navigate in practice. Most of the difficulties according to Maxwell arose where people failed to alert them in time. He recalls 'the ones that were rather difficult to handle were ones that sent it in and said, “We're sending this in an hour's time” to somewhere or other and we had to think a bit about how to handle that appropriately' because 'there was a major commitment to making the information as widely available as possible'. COG-UK took the position that authors acknowledge the consortium when using its data (Maxwell transcript).