An open and transparent process to select ELIXIR Node Services as implemented by ELIXIR-UK

Introduction

ELIXIR, the European infrastructure for life science data¹, is made up of individual Nodes, one for each of the organisation’s constituent members (20 as of 1st November 2016: Belgium, Czech Republic, Denmark, EMBL-EBI, Estonia, Finland, France, Germany, Ireland, Israel, Italy, Luxembourg, Netherlands, Norway, Portugal, Slovenia, Spain, Sweden, Switzerland and the UK), and a coordinating hub. The individual ELIXIR Nodes provide the services and resources that support the five pillars of ELIXIR (Compute, Tools, Data, Interoperability and Training infrastructures).

ELIXIR nodes need to be able to evolve their contributions to ELIXIR by bringing new services and resources. ELIXIR identifies two types of service: Node-funded services, which are funded nationally and are contributed to ELIXIR from a national Node; and Commissioned Services, which are funded by ELIXIR as a whole via the ELIXIR Hub. In some ELIXIR Nodes, Node-funded services receive funds through their national Nodes; in the case of the UK’s Node, ELIXIR-UK, resource funding is through direct grant funding to resources and services from the national funders. In ELIXIR terms, these are still labelled as “Node-funded”. The process described in the present article was set up to identify Node-funded services and resources for ELIXIR-UK. ELIXIR sets high standards for the services it provides. Consequently, nodes need to take full account of these requirements when selecting and proposing their services, which are ultimately judged for suitability by the ELIXIR Scientific Advisory Board (SAB) and Board of ELIXIR (see the online ELIXIR Handbook for more detail).

ELIXIR-UK was established in September 2013, and as its first contribution to ELIXIR took on a thematic focus, namely of coordinating training activity. More recently, it has sought to expand its remit. To address the SAB’s recommendation that Nodes put in place “mechanisms to ensure that each Node continues to be representative of the Bioinformatics efforts within the country”, ELIXIR-UK developed a process to choose new services and resources to add to its existing portfolio. Its aims in developing this process were to:

Process overview

As illustrated in Figure 1, the process implemented by ELIXIR-UK went through seven key phases, which are expanded on in the following sections:

Figure 1. Summary view of the information flow in the ELIXIR-UK Node Expansion process.

Items in the red correspond to the numbered list of key phases. Numbers in red correspond to the seven phases of the process listed under the section “Process overview”.

Strategic prioritization

The requirement to ensure that each Node continues to be representative of the Bioinformatics efforts within the country could be seen as open ended, and thus could ultimately lead to an ill-focussed collection of resources and services. To avoid this, ELIXIR-UK identified a set of priority areas within which to focus submissions to the process. These were initially identified by discussions within the Node and were refined by discussion with the Node’s funding organisations (Biotechnology and Biological Sciences Research Council [BBSRC], Medical Research Council [MRC] and Natural Environment Research Council [NERC]) and with the Scientific Development Group (SDG), which was a community body set up by the Node that is tasked with identifying new node resources (see below).

As a consequence of these discussions, Expressions of Interest (EoIs) were invited in the following priority areas, identified as being of high strategic importance within the UK:

Identifying possible candidates

ELIXIR-UK aimed to reconcile two potentially conflicting drivers in developing its expansion process. Firstly, it wanted to be as open to the UK bioinformatics community as it could. This is an ongoing challenge because a) ELIXIR has incomplete brand recognition within the UK community, and b) is not well-regarded by some, being seen either as a closed club or unproductive. Secondly, the Node wanted to ensure it received Expressions of Interest from potential services and resources that were demonstrably of high value to the international life sciences community. To address these requirements, the Node approached the recruitment of potential candidates in two ways. Firstly, it publicised its “Node Expansion” process well in advance using its web site, Twitter and word-of-mouth. Secondly, it sent targeted emails to potential candidate resources. These were identified using a variety of inputs:

The Scientific Development Group

The key body in the Node’s expansion process was the SDG. This was set up by the Node to evaluate EoIs to join the Node against a set of published criteria (see below). This group was also involved in refining those criteria and providing suggestions of resources to be invited to provide EoIs.

The membership of this group was based on suggestions from within the Node and from its funders. The group’s composition reflected the priority areas identified for the expansion, geographic spread, and the inclusion of at least one industry and at least one overseas representative. The Chair was chosen for his experience as a senior officer of a UK funding agency and knowledge of appropriate processes for activities of this kind. For the record we note that the group did not have an appropriate gender balance (it was 100% male). This is a defect we intend to remedy in future.

Assessment criteria

Over time ELIXIR has been evolving both its classification of resources and its criteria for selecting them. During the period of the UK Node’s expansion process these definitions and criteria continued to evolve. The assessment criteria developed by ELIXIR-UK were developed by internal discussion and in discussion with the SDG and were also discussed informally with the leaders of Work Package 3 of the EXCELERATE programme (Jo McEntyre and Christine Durinx), as their criteria developed in parallel. The final set of criteria, which were provided to applicants as an openly shared Google document, were:

To facilitate applicants demonstrating that their resources fulfilled these criteria, an Expression of Interest template form was provided, also via Google documents.

The criteria developed by EXCELERATE Work Package 3 have subsequently been finalised and form the basis of the ELIXIR process for selection of Core Data Resources².

Engaging the community

As outlined, it was important to ensure community buy-in to this process (in order to ensure that the Node was able to engage sufficient high quality resources) and at the same time it was important to be sure that community members who might be interested in participating in ELIXIR-UK were aware of what was required and the expectations that would be placed on them as ELIXIR-UK Node resources. Formal community engagement took place in two phases: a webinar, led by the Head of Node (CAG) and Node Coordinator (JMH), in February 2016 and a workshop, hosted at the Wellcome Trust building in London, in March 2016. The aim of the webinar was to introduce ELIXIR and ELIXIR-UK and the rationale behind the node expansion process. The aim of the workshop was to introduce and discuss the assessment criteria in detail, so that potential applicants could be clear as to what was required. The presentation given at the workshop is available via Slideshare. At this stage a deadline was set for the receipt of Expressions of Interest by the Node. It is worth noting that the deadlines for the process were tight: EoIs were requested by the end of March 2016 and the assessment meeting took place at the end of April with some iterations taking place in May. We were fortunate in being able to run such a tight schedule due to a) clear and lightweight requirements for the EoIs; b) what we believe to be clear and effective communications; and c) motivated applicants and SDG members.

Assessment of Expressions of Interest

EoIs were assessed by the SDG against the published set of criteria. To facilitate assessment of EoIs, three group members were allocated to each EoI (18 were submitted). The three members were asked to score EoIs from 1 to 4: 1 = ready for inclusion in ELIXIR (“infrastructure ready”); 2 = further discussion or clarification needed; 3 = not ready, but suitable enough to be placed on a roadmap for future inclusion; 4 = not suitable. The assessments for each EoI were introduced by one member of the group leading on to an open discussion. Representatives of the Node funders and the ELIXIR-UK executive observed the meeting to give advice on strategic alignment. EoIs were given a consensus final score using the same scale as before, with a score of 2 in this case representing the need for further clarification of issues raised by the group. Resources given a 2 rating were asked for further information, which led to their final score being revised upwards or downwards in a subsequent iteration.

Results of the assessment

The outcomes of the assessment are summarised in Table 1. Nine EoIs were considered to be infrastructure ready (score of 1) on the first pass of assessment, and a further five were asked for more detail on their proposal (score of 2).

An iteration of discussions with resource scientists allowed questions raised by the SDG to be considered further. Where these were answered satisfactorily, resources were moved up to infrastructure-ready status. Otherwise they were put on the roadmap or, in one case, referred to the ELIXIR SAB for further comment (in this latter case, SAB guidance subsequently resulted in it being accepted as infrastructure ready).

After ratification by the ELIXIR-UK executive and notification to the ELIXIR Hub, highly rated resources were included directly into the Node’s portfolio and were included in the Node Application presented to the ELIXIR SAB in June 2016 and the ELIXIR Board in November 2016. Others were placed on the Node’s roadmap for possible future inclusion.

The services and resources selected as ready for immediate inclusion are listed in Table 2.

Future activities and conclusions

We believe that the process outlined here was open, transparent and fair. We note that the “success rate” of the process was high. No resources were rejected outright and more than 70% were promoted immediately to the Node’s portfolio. This does not reflect a lax process, but is likely to have had a number of contributing factors, including:

Another aspect of the process we outline here is the short time period over which it was carried out. In particular, resources were only given four weeks to submit EoIs. A number of features of the process facilitated this: clear timelines, clear guidance as to what was required, the availability of a template for EoIs that helped proposers to compile their EoIs, and lightweight requirements for completing EoIs, which were nevertheless sufficient to allow the SDG to carry out its work effectively. Engagement at a senior level by both the Node and proposers was also important. It was also important to organise meetings, especially of the SDG, sufficiently ahead of time to allow members to both assess the EoIs and attend the meetings, either in person or remotely.

To maintain and continue to improve the Node’s alignment with UK research strengths, it plans to hold regular refresh exercises to introduce new resources into the Node. Plans for how this will be done are currently under development. To pursue this process we expect that we will need to develop community engagement in the specific priority areas, so that potential proposers are primed.