The UK Biobank’s 2019 scientific conference concluded this month, showcasing fantastic speakers on topics ranging from use of the UK Biobank to using genomics to understand individual risk to strategies for health outcome phenotyping. Among these speakers was Professor Anthony Whetton, Professor of Cancer Cell Biology at the University of Manchester and Director of the Stoller Biomarker Discovery Centre. FLG spoke to him about his talk on large-scale multi-omic data within UK Biobank, and about its effect within the clinical space.

FLG: Could You Tell Us More About Your Talk at Biobank 2019?

AW: In my talk I was charged with conducting an analysis of the proteomic and metabolomic strategies for UK Biobank. I began by considering the repository: what amount of biological material was available, and how a multi-omic approach would impact the amount of material stored. The answer was that using these new techniques, you wouldn’t really dent the amount of material stored in UK Biobank to any great extent.

From there we moved onto the question of how to analyse this material. Already the UK Biobank has analysed a series of biochemicals which have biomarker credentials, so to speak. So there’s a series of tens of biomarkers that have already been assayed. These are what I call the “known knowns”. Then there are a host of other biomarkers out there that aren’t being assayed due to cost restrictions and other factors, the “known unknowns”. We need to assay those with proteomic and metabolomic approaches. And finally there are the “unknown unknowns”, and we can identify new biomarkers using techniques like proteomics and metabolomics. In the case of metabolomics we’ll use nuclear magnetic resonance (NMR) and mass spectrometry; for proteomics the best technique is mass spectrometry.

Then we need high-end informatics, integrated with clinical records, to first understand this information in its own right, and then undertake the multi-omic integration process – which we can do using the kind of techniques found in a big data institute or leading health informatic centre like Manchester.

FLG: Why is it Important to Have Multi-Omic Data in the UK Biobank?

AW: It’s widely accepted that the UK Biobank is a unique resource, but it’s also one established with great foresight. This foresight included the storage of samples such as plasma, saliva, and urine, from which we can obtain a lot of biochemical information in respect of identifying determinants of risk of ill-health and general wellness.

The best way to do that kind of work these days is not by measuring a single molecular species at a single time, which is wasteful in terms of the amount of sample used. The best way is to use -omic strategies, the principle ones in this case being proteomics and metabolomics. With relatively little material, that way you can get information on hundreds of thousands of compounds and proteins.

The other reason this is useful is the amount of genomic data we have from volunteers at the UK Biobank. Marrying the proteomic data to the genomic and metabolomic data, and to the healthcare records and other volunteer info from UK Biobank, will give us a huge opportunity to identify the relationships between genome and metabolome, genome and proteome, and aspects of determination of risk.

As we’ve seen with the first release of exome sequencing data from UK Biobank, you can identify genes and their associations with specific diseases already, and that can be further extended with the proteomic and metabolomic work.

FLG: Beyond UK Biobank, How Close is Multi-Omics Data to Full Academic Integration?

AW: In terms of proteomics, the Medical Research Council funded the establishment of the Stoller Biomarker Discovery Centre in Manchester so we could use mass spectrometry-based approaches to identify and quantify biomarkers of value in cancers, and in inflammatory and other diseases. So there are facilities for that. The chief facility, because of size and scope, is in Manchester.

In terms of academia and metabolomics, the National Phenome Centre, the University of Birmingham and others can perform metabolomics assays to an appropriate level. And so there’s the ability for clinical researchers to tap into those research bases, both for their own cohorts and for the development of analyses of major cohorts like the UK Biobank.

FLG: Since That’s the Case, Would You Say Multi-Omics is Successfully Integrated at this Point?

AW: It’s been a big challenge to integrate clinical records with -omics data, and that includes proteomics data. There are health informatician groups around the country capable of doing this work, and Manchester certainly can achieve integration with respect to the analysis of proteomic plus nucleic acid-based -omic data, and we are in a very good place regarding integration with clinical data and getting good high-quality information using new techniques and approaches like topological data analysis and latent class mixed model methods, in order to get as much as we can out of an -omics data set.

FLG: We’ve Already Spoken About the Big Challenges Around Data and Cost. Are There Any Others Preventing Full Multi-Omics Integration?

AW: Cost and data are certainly the major challenges. There are others associated with cost: this data needs quality assurance and control, done at a high level, which is costly. After that there is the big issue of how to use data effectively.

Quality and consistency of collection is always going to be an issue, of course, and storage too. You need to store all the samples under the same conditions or you could end up with artefacts. Luckily, as I already mentioned, the amount of material required for these assays in respect of proteomics is really low, and in respect of metabolomics is relatively low, so I think that at least isn’t an issue in respect of biobanking globally.

FLG: Obviously Multi-Omic Data has a Lot to Do with Precision Medicine. How Important is Multi-Omic Data to Precision Medicine, and How Will it Influence it in the Future?

AW: The big issue around precision medicine is treating the right patient at the right time with the right drug. We need biomarkers to do that, indicators of when we should be treating the patient and what with.

The second aspect of this is that once you’ve given a drug to a patient you need to determine if they are responding or not, so that theranostic area needs to be considered too. You have to collect samples after a drug’s administration and see whether there’s a response or not.

So in respect of the -omics strategies being deployed, we currently don’t have enough biomarkers to stratify patients for the right treatment, and certainly not enough to determine early on if there’s a response to therapy so we can take informed decisions about further treatment strategies. So I think these are areas where -omics can impact precision medicine positively.

FLG: Is There Anything Else You’d Like to Add About Your Talk?

AW: The opportunities UK Biobank offers for us to understand wellness and ill health are fantastic. I am really enthusiastic about bringing techniques like proteomics to UK Biobank, because everything else is so well set and thought through, and is such a stable platform for -omics enterprises to build upon.

Look out for more interviews from Biobank 2019 coming soon.