Technology Touching Life Consultation
From BioImagingUKWiki
Background
(This text is from the instructions. BioImagingUK has been asked to provide responses to this consultation.)
Building on our track record of working together to support interdisciplinary research across the physical, life and biomedical sciences, BBSRC, EPSRC and MRC are at an early stage in scoping a joint strategy to foster more diverse, fundamental, interdisciplinary technology development research. This is a new theme provisionally titled ‘Technology Touching Life’ (TTL). As part of the initial scoping exercise this consultation is being sent to universities/institutes, individual researchers, learned society contacts and industry representatives to garner views from the scientific community in order to inform the development of the TTL theme.
Discussions on TTL across the three Research Councils were initially stimulated by the EPSRC report The importance of engineering and physical sciences research to health and life sciences, published in May 2014. Fundamental breakthroughs in the life and biomedical sciences are often based on new physical science-based research technologies, which in turn often open up longer term opportunities for the economy and society. The TTL strategy aims to stimulate and support interdisciplinary collaborations to explore novel technologies and approaches that address application-driven challenges. By enabling joint working and two-way flow of ideas between life scientists and engineers/physical scientists we expect that TTL will ensure the UK leads future waves of foundational technology discovery for the life and biomedical sciences, and create new opportunities for commercial development.
Questions and Submitted Answers
Contributors (alphabetical order): Luc Bidaut (Dundee), Ilan Davis (Oxford), David Fergusson (Crick), David Gadian (UCL), David Hawkes (UCL), Boguslaw Obara (Durham), Jacky Pallas (UCL), Jason Swedlow (Dundee), Theresa Ward (LSHTM), Ann Wheeler (Edinburgh), Ray Wightman (Cambridge)
Within the broad scope of Technology Touching Life:
Q1. What are the ‘sweet spot’ areas where there is high potential for closer alignment across physical sciences and life sciences to lead to major advances?
There are literally dozens of opportunities across the life and biomedical sciences where a close alignment with the physical sciences could lead to major advances and discoveries. Three of the most pressing opportunities involve genomics and proteomics, imaging and the data sciences.
- In genomics and proteomics, we are in the middle of a revolution. Delivery on the potential of these technologies requires innovations in chemistry, detection, automation and data analysis— in almost all cases, innovative technologies or capabilities are needed that are smaller, cheaper and faster. These should be considered major strategic scientific, social and commercial priorities given the potential impacts on UK and world society. This revolution will occur on three scales:
- Single Cells. The arrival of in situ transcriptomic and proteomic profiling will allow a full molecular profiling of a living cell, through the cell cycle, various steps in differentiation and in response to environmental stressors (e.g., toxins). Critically, these technologies can be combined with spatially and temporally resolved imaging, to reveal not just which molecules are present in a cell, but where they are and what they are doing.
- Individuals. With new advances in DNA sequencing and proteomic technology, most individual humans, domesticated animals, and crops will soon have their genomes sequenced, and their gene expression and protein abundance and modification profiles mapped and known. The impact of having, for example, the gene expression and protein abundance profile of many different tissues in an organism or the full lifecycle of an important food source won’t be clear for several more years. It is very likely that this data, when combined with other information (as described below) will lead to important fundamental discoveries, diagnostics, and new therapeutic approaches for modern diseases. Personal DNA sequencing is potentially a quite lucrative commercial market. Genomic profiling is, however, just the beginning. Transcriptomic and proteomic profiling will report on an organism’s response to their environment, and provide molecular insights into health, diversity, and pathology in all living organisms.
- Populations. The accumulation of genomic and proteomic datasets on the scale of communities and nations is beginning to transform healthcare delivery and several other domains— insurance and criminal law are just two examples. The Genomics England 100,000 Genomes initiative (http://www.genomicsengland.co.uk/) is one example of a public-private partnership to deliver population-scale genomic data for the benefit of health delivery. To truly achieve the diagnostic and predictive value of these genomic datasets, they must include other biomolecular profiling, to include comprehensive measurement of transcription, protein abundance and modification, metabolism and these must be placed into spatial and temporal contexts— through imaging.
- Imaging measures the molecular and structural composition and dynamic behaviour of biological and biomedical entities. Most imaging modalities are capable of measurements in space and time, providing a view into the structure and dynamics of biological and biomedical systems. An ongoing challenge across all imaging modalities is to deliver spatial and temporal resolution on scales that match the scientific, diagnostic and therapeutics needs. This explains the importance of recent improvements in resolution in light, electron and medical imaging— improved resolution allows researchers and physicians to “see” and measure more carefully and more accurately.
The ubiquitous use of and critical need for imaging to study all aspects of microbial, plant, animal and human biology cannot be understated. Because imaging provides spatiotemporal maps of the physical structure and chemistry of biological and biomedical systems, it will contribute to several of the major strategic initiatives in the UK. Antibiotic microbial resistance, food security, health and social impacts of an ageing population, improved treatments for cancer, and an understanding of the function and diseases in the brain will all be impacted and advanced by new capabilities in imaging-based measurement, modelling, diagnosis and therapeutic intervention.
By combining the physical, life and biomedical sciences we can foresee the development of new imaging modalities that reveal chemical and physical structure and dynamics that have to date not been available. The potential impact of these advances has recently been summarised in EPSRC’s ‘Healthcare Technologies Grand Challenges’’ (http://www.epsrc.ac.uk/research/ourportfolio/themes/healthcaretechnologies/strategy/grandchallenges/). This and several other reports, including BioImagingUK’s previous strategic prioritisation have highlighted the importance of capabilities to combine imaging and other modalities across scales, e.g., to measure and localise transcriptome, protein and metabolite abundance in cells at nanometre resolution, and combine this information with 3D measurements of cell and tissue architecture, where possible in the intact, living state (a microbial biofilm, a plant, animal or human). Achieving these goals in the 21st century depends on the development of new probes for revealing molecular and structural composition, new detection modalities for revealing the composition of biomedical systems and new data integration and analysis tools. Some examples of imaging “sweetspots” that can be considered as components of the wider multi-scale challenge:- Mass Spectrometry Imaging. There are several labs developing imaging using mass spectrometry where isotopic analysis is used to reveal the chemical makeup of biomedical systems. Full development of mass spectrometric imaging (MSI), to include labeled probes labeled with specific mass markers could revolutionise the mapping of biochemical species in cells and tissues.
- Biological Imaging. Super-resolution techniques such as Structural Illumination microscopy, Single Molecular Localisation Microscopy (e.g. PALM, STORM, GSDIM) and Stimulated Emission Depletion microscopy technologies are prime areas for new developments that probably can only be achieved by alignment between the life and physical sciences. In many respects these tools are in early stages of development for application in biomedical research. The technologies are considerably slower, more cumbersome, lack good probes and analytical tools. Development of correlative methodologies to utilise two or more imaging technologies, in particular electron and atomic force microscopy with other light based modalities requires considerable development. Marrying these techniques up would yield valuable new data however only close development across physical and life science disciplines can allow for breakthroughs in these area.
Biological imaging is now progressing to larger scales, enabling imaging of embryos and small whole organisms at high resolution. Light Sheet Microscopy, Optical Projection Tomography, small-bore Magnetic Resonance, micro-CT and Fluorescence Molecular Tomography are all established techniques that reveal the molecular composition and/or architectural details of tissues or small-organisms, but require improvements in probes, resolution and speed. - Medical Imaging. Advances in physics, modelling, and image-guided planning for surgery and radiotherapy can improve precision or surgical intervention, leading to fewer side-effects, faster recovery, and better outcomes. As just one example, the investigation of tumour metabolic status using novel magnetic resonance hyperpolariser technology could prove to be of enormous value in helping to guide medical and surgical management. Again, the use of multiple modalities, to correlate modes of contrast across different size and resolution scales will produce insights into the biochemical and physical basis of disease in the context of a whole organism.
Developments in medical imaging have long been achieved by close collaboration between biomedical scientists, clinicians, physicists and chemists and serves as one example of the power of interdisciplinary science.
- Data Management, Analysis and Visualisation. All of these trends move towards creating larger and much more complex datasets. The computational requirements for handling the data created by genomic, proteomic and imaging studies is growing rapidly, and most estimates indicate that data volumes and the need for data processing will exceed the computational and storage capacities available [Nature,498,2013]. A critical need in the life and biomedical sciences is the delivery of new data compression technologies, data management systems, analytic and processing methods and tools, and complexity reduction algorithms that work at large scales to deliver the most optimal results that enable new discoveries, diagnostics and therapies.
- Data Management and Access. Platforms for managing access, analysis, and sharing of the datasets collected using the technologies described above are emerging as critical bottlenecks for delivering maximum value, and for complying with Research Council requirements for data provenance and publication. Most biological data is shared via download, via FTP or HTTP. As dataset size increase to the multi-GB - TB scale and beyond, data download will become impractical. Virtual or cloud-based solutions with defined application programming interfaces (APIs) can be built to enable remote data access and virtualised analysis.
- Data Integration. While individual domains— genomics, proteomics, imaging— are developing tools for handling their datasets,capabilities for linking these different datasets are still being developed. For example, linking biomolecular (genomic, transcriptomic and proteomic) profiles of single cells with measurements of cell and tissue shape, structure and other physical parameters will be critical for understand the molecular basis of homeostasis and disease.
- Machine Learning. Application of machine learning and deep learning (DL) methods to biomolecular and imaging datasets is now established. New DL methodology has the potential to revolutionise the definition of classifiers and other discrimination tools to automatically identify and discern phenotypes at various scales. DL applications on a wide variety of 2D image types and content now deliver impressive results (http://arxiv.org/abs/1409.0575); implementing these approaches in the context of multi-dimensional biological and biomedical images promises to deliver effective image-based search and retrieval.
These are the most obvious big wins for combining efforts in the physical and life sciences. Application of new technology inevitably reveals opportunities and knowledge that was simply unimaginable before its development. One example is the size and importance of the microbiome in multicellular organisms and the interaction of an individual animal, plant or human with its own microbiome. It is nonetheless critical to invest in the development of these technologies so that the UK can benefit from the scientific, social and economic benefits they will create.
Q2. What trends or developments in engineering and physical science technologies are already emerging which you think will have impact in the life and biomedical sciences?
There are several ongoing major efforts in the physical sciences that will have massive impacts in the life and biomedical sciences across research discovery diagnostics and therapeutics. Some major points that we foresee will be extremely important:
- Efforts across the range of photonics to make increasingly smaller, brighter, and cheaper light sources should translate into research, diagnostic, and therapeutic tools that have a wide range of applications.
- The development of new probes for measuring chemical and structural composition and environment in living tissues and organisms is proceeding and will be an important part of future the development of new discoveries and diagnostics. Moving forward the development of multi-modal imaging probes where complementary imaging modalities can be used to reveal the biological and biomedical structure and dynamics will be incredibly important for understanding the molecular basis of disease and the effect of candidate drugs on these diseases.
- Nanotech, in particular small engineered materials, which are useful for delivering medicines to specific locations or under specific conditions should be an extremely powerful technology that will deliver benefit for society and science.
- A huge opportunity for the development of computational methods and technologies for managing, processing, analyzing and understanding the very large, heterogeneous ("big data") datasets acquired across all domains of the life and biomedical sciences will be fundamental for delivering the promise of stratified medicine and healthcare.
Q3. What important areas in the life and biomedical sciences are currently limited by existing technologies and require new technology developments to ‘unlock’ discovery opportunities and deliver a 'step change' in understanding?
Across all life and biomedical sciences there is a common challenge for handling, sharing, processing, publishing, and delivering knowledge from the large, multidimensional, heterogeneous datasets. This is now a familiar challenge in all domains of the life and biomedical because of the great advances in automation and detection in genomics, proteomics, imaging, and other quantitative modalities. As important as these various data acquisition technologies are, the computational methods and tools for handling these datasets has not kept up. Dedicating resources for computational methods and tools is a critically important opportunity for investment and collaboration across the physical and life sciences.
Q4. What do you see as the key features of successful models of close working between physical sciences and life sciences in fundamental discovery research? Please use specific examples, national or international, where possible. You views on unsuccessful as well as successful models are welcome.
The current drive for interdisciplinary science is laudable and correctly recognizes the need to bring expertise from many different domains together to achieve progress on strategic goals important for science and society. However the current career and recognition models derive from the now outdated need for an individual scientist to prove his or her own success for career progression, recognition and reward. While it is possible to bring scientists from different domains with different recognition and career progression mechanisms together, merely pretending these will combine and easily meld together is naive.
There are examples where true cross-disciplinary collaboration has occurred and the normal academic career progression has been discarded. One notable positive example is the Allen Brain Institute (http://www.alleninstitute.org/) where engineers, biologists, and informaticians have collaborated to build (over a 12 year timescale, a timeline that stretches long past conventional project horizons) the Allen Brain Atlas, a resource that now is becoming the foundation for the development of new scientific discoveries in neuroscience. One notable aspect of the Allen's work is that career progression and success for the Allen Insititute’s scientists are solely based on their contribution to the Allen Brain Institute’s mission and not their individual publication record, citation statistics, etc. Thus they are measured based on their contribution to the interdisciplinary project, not what part of the project they have somehow retained or branded for themselves. One message from this story is that large interdisciplinary teams that retain a focus on individual success may not be able to achieve the scale of accomplishment, discovery and contribution they might otherwise simply because the individual team members have to focus on their own progression and success.
Q5. What structures, activities and mechanisms help establish a culture of interdisciplinary research and strong interdisciplinary leadership in universities, institutes and centres?
Academic Consortia. The best and most effective way to establish a culture of interdisciplinary research is to promote the formation of active research-driven consortia that are formed by PIs and other researchers to address or develop research programmes and proposals that respond to a specific opportunity. Strategic opportunities can be driven either by universities or Research Councils with pump priming funding or strategic calls. Different scales can be used to drive the formation of interdisciplinary consortia in different domains or with different scales of challenges. The specifications for these consortia should be relatively loose: a specific formula for proportions or types of biologists, physicists, chemists, engineers and/or computer scientists is usually too prescriptive to allow the research-driven definition of a well-constructed team focussed on a specific problem. Strategic research council funding initiatives can prime the pump for the development of these consortia especially if they include directives for building interdisciplinary teams. Good examples are the Next Generation Optical Microscopy call and the UK Regenerative Medicine Platform.
In 2015, there are now several examples of interdisciplinary consortia operating at and across UK universities. These operate using many different models, but a common theme is the linking of scientific activities and disciplines enables opens the doors to scientific and funding opportunities that aren’t available otherwise. In this sense, UK scientists have already formed interdisciplinary consortia across several themes and are ready to respond to the funding opportunities presented by the Research Councils.
Careers. At the level of the university there must be mechanisms to recognize the commitment of research time to interdisciplinary programmes that stretch the conventional research portfolio within each department. If a physicist collaborates with a life scientist and the ultimate output is a paper in a biomedical journal then the contribution for the physicist who appears as a middle author on the paper must be valued in his or her home department.
Industrial Partnerships. There are huge opportunities for linking academic and industrial efforts across the life sciences, biomedical research, engineering and physical sciences. Examples include the commercialisation of high throughput sequencing (http://www.illumina.com/technology/next-generation-sequencing/solexa-technology.html) and efforts to solve the computational challenges in personalized medicine (http://www.intel.com/content/www/us/en/healthcare-it/collaborations/ohsu.html). Again, the output of such partnerships may test the normal criteria for academic progression, but are an important component of a modern university's activity.
Q6. Please comment of the role that effective and inspirational leadership plays, giving examples where possible.
A critical component of interdisciplinary projects is strong leadership and management. Ensuring that all partners in an extended collaboration are interacting as necessary, delivering on promised tasks and committed to the promised timelines milestones and goals is an absolute necessity. In addition, there must be a respect of the needs of different project members, whose career goals (clinical progress, academic tenure, commercial success) will often not overlap.
Defining ‘inspirational’ leadership is difficult. The most inspirational leadership sometimes comes from supremely competent managers-- the inspiration comes from the consortium and collaboration delivering its scientific and strategic goals. Laissez-faire management is often a major reason why strategic projects fail.
Q7. How might the Research Councils working with research organisations (e.g. universities and institutes), industry and other stakeholders help address the issues raised in response to the questions above? Please provide examples of successful approaches (both national and international) where applicable.
The first priority in considering strategic options for driving interdisciplinary science must be the science and technology development itself. The funding and delivery of the best and most important advances in science, diagnostics and therapy must be the overriding goal. Collaborations between the physical and life sciences are then a means to achieve this.
Academics are generally very good at bringing together teams when the correct challenge is set. They can act as the glue between clinic, research and industry and thus enabling such people to bring together and then apply for funding, which may have many traditional funding routes (eg MRC, NHS, EPSRC, Innovate UK) in one go is important. This might be based upon an initial pure research project with then a seamless transmission to application and commercialisation (TSB, Innovate UK). This progression will always be challenging, but several funding mechanisms target specific parts of these steps (e.g., Pathfinder and several TSB schemes).
One of the challenges in delivering these collaborations and thus their potential scientific output and impact is developing the productive collaborations in the first place. Building bridges between different disciplines and establishing strong, strategic collaborations takes time and resources. It is therefore worth considering funding interdisciplinary collaborations at different levels. Shorter term proof-of-concept projects should be separated and funded separately from longer-term strategic efforts. The research councils aligned with universities, potentially using similar mechanisms as the MRC Confidence in Concept schemes could deliver these tiered funding lines.
Q8. Finally, we would welcome any other comments you have on developing the 'Technology Touching Life' theme.
TODO