Integrating genomics into healthcare is a monumental task that needs to be broken down. Here, Dr Ang Davies, Lecturer in Healthcare Sciences at the University of Manchester, and Dr Jan Taylor, Clinical Scientist in Bioinformatics at the St. James Institute of Oncology, use their combined expertise to explore the challenges faced by clinical bioinformaticians in the NHS. It’s an exciting time at the coalface, and there’s still a lot of work to be done.

Genomics is beginning to have a significant impact on areas of medicine other than clinical genetics, including cancer, cardiology, microbiology and ophthalmology to name but a few. The realisation that we can now very quickly and relatively cheaply sequence a number of genes simultaneously that will influence a treatment regime, provide a diagnosis or even exclude certain diseases/conditions means that clinicians are opting for this route rather than conducting a sequential genetic testing which can be time-consuming and often does not lead to a diagnosis – often referred to as the diagnostic odyssey. What has not been addressed to date is the efficient management of huge quantities of complex genomic data that sequencing technologies produce and the shortage of skilled clinical bioinformaticians to analyse and interpret this data.

There is a small cohort of clinical bioinformaticians working within the NHS in the UK, who come from diverse backgrounds. Many have come from a biological or computer science research or translational research background, whilst others were trained as a clinical scientists and have developed their own bioinformatics skills within their clinical centres. Bioinformatics is a broad discipline, covering a wide range of specialisms and current clinical bioinformaticians have very different backgrounds which impacts on their mode of practice. This could be likened to a surgeon for example, whilst they might all be described as surgeons clearly there are vast differences between the training and role of an eye surgeon to a heart surgeon. The use of programming languages that a bioinformatician might use in their work is a good example of this, varying from Python, to Perl, to Java and so on.

In the UK, Modernising Scientific Careers led by Health Education England has gone some way to addressing the requirement for training focused on clinical need, leading the development of a new Scientist Training Programme in Clinical Bioinformatics. This training comprises 3 years of work-based training and a parttime masters programme undertaken at The University of Manchester, enabling them to apply for clinical registration through the Health and Care Professions Council. Unfortunately however the first cohort will not finish until July 2016, in the meantime there will be the opportunity for those currently practising as clinical bioinformaticians to undergo a process of equivalence to this programme to also become registered scientists and therefore also progress to Consultant Clinical Bioinformaticians. These leadership roles will be vital in years to come if we are to truly benefit from this genomic revolution.

So what is the situation really like at the coalface for current bioinformaticians managing, analysing and interpreting all of this genomic data in the healthcare setting? In practice the experience is vastly different, varying from Trust to Trust, and even between departments of the same Trust. Roles vary from being part of a team (3-4 people), or a single role embedded in a clinical department. In an ideal world one might imagine a team of bioinformaticians made up of a system administrator, programmers, and web developers, although most centres are a long way from this currently. However with the introduction of bench-top Next Generation Sequencing instruments most centres, be they large or small, do have access to the technology and so most likely will be analysing some kind of genomic data requiring the development and implementation of a bioinformatics ‘workflow’ to analyse this data. This workflow will enable the signal generated from the sequencer, to be converted into sequence, reassembled and mapped back to a reference genome and then analysed for differences which are then further analysed using specialised tools and databases. This might be done using bespoke ‘in-house’ bioinformatics workflows, commercially sourced workflows, or a hybrid of the two. The workflow choices depend on many factors; choice of sequencing machine, DNA capture method, sample type, available computing infrastructure, and previous bioinformatic experience. This makes clear the difficulty in standardisation of standard operating procedures and protocols between laboratories currently and this is unlikely to change, as departments have different priorities for procuring NGS capabilities (budget likely being the overriding influence). This therefore places the emphasis and importance on the need for validation against test data sets and adherence to appropriate recording of programming code and version control, and databases and tools that are used in the bioinformatics analysis. Pressure is being brought to bear in this regard with the introduction of ISO standards that bioinformatics workflows that will need to be adhered to for clinical laboratories to retain accreditation.

Provision of the appropriate IT infrastructure required to store and analyse genomic data is also an issue which needs to be addressed. As with sequencing technologies, the available computing power to a bioinformatician varies greatly, from a single standard workstation, a linux server through to a managed HPC facility, depending on the centre, and sometimes the available links to University research computing facilities.

The landscape of IT provision in the NHS is complex and each hospital Trust is likely to be different, therefore the bioinformatician has to communicate with many stakeholders to ensure this functions as required: this includes data storage; health informatics and governance, who may have important information that links to part of a patients records; clinical colleagues – this will be discussed in more detail; data communications – that look after the networks through which these large quantities of data are moved around the hospitals; IT services; and also research IT services – sometimes it is easier for hospitals to collaborate with universities to analyse this kind of data as they already have the systems in place; and the other important route of communication is with other bioinformaticians to try to establish best practice in this developing discipline. Dealing with these complex structures requires tenacity, perseverance, leadership and strong communication skills.

It is important for the bioinformatician not to lose focus of the fact that there is a patient, with most probably additional family members in the case of inherited diseases for which the outcome of their work could have a huge impact. At this moment it is unlikely for the bioinformatician to directly see a patient but certainly in the larger centres they will take part in multi-disciplinary meetings with clinicians, senior clinical laboratory scientists and also genetic counsellors, where they will be helping to inform a clinical diagnosis or treatment regime. Therefore they have a really important part to play in educating their clinical colleagues such that the workflows and bioinformatics steps are more transparent and stand up to the necessary rigour that other laboratory processes are required to achieve. One of the most important roles that the bioinformatician needs to undertake is to ensure that patient data, including symptoms and phenotype are captured by clinical colleagues and linked with associated genotype. Achieving this on a much larger scale with National/ International databases such that more data and information can be shared subject to patient consent within the clinical genomic community means that the patients particularly with rare conditions and diseases will benefit more quickly from advances in genomic sequencing. It also fundamental that the bioinformatician is informed of any changes to laboratory protocols such as the addition of additional genes or changes to protocols that will impact on the bioinformatic workflow such that these can be planned for, validated and tested prior to being relied to inform clinical diagnoses and decisions.

Clinical bioinformatics is a newly recognised healthcare profession, and as such there is a lot of work to do to increase its reputation in the healthcare setting. For those of us working at this frontier, it’s a time for building a community of best practice, sharing a diverse set of skills and experiences, and embedding bioinformaticians as integral members of the multi-disciplinary care team. 

This article first appeared in Issue 3 of Front Line Genomics magazine.

More on these topics