Realising the Potential of Genomics through Connection

A couple of weeks ago I had the joy of attending the Festival of Genomics and Biodata. It was the most diverse conference I have attended thus far, bringing together academic researchers, clinical professionals, biotech companies, patient groups, cloud providers, and those collecting genomics data. The variety of sessions I attended really broadened my understanding of where genomics is today. Here, I highlight some of the themes that came across throughout the festival, which broadly centre on the idea of connection.

Connecting Biotech to Opportunity

As an example of connection, many biotechnology companies and their academic partners have agreed to pre-competitive collaboration in genomics. These agreements are acknowledgements that more progress can be made in genomics through collaboration than would be made if each organisation tried to perform research independently. For example, genomic data is more powerful in larger quantities, so it makes more sense for everyone to co-operate on the data gathering front and compete in product development instead. Additionally, insights generated from –omics data are publicised, even when they have been generated in-house by a biotech company, because collective –omics evidence is more useful than dribs and drabs.

Pharmaceutical companies in particular value access to genomic data because evidence suggests that genetic validation for drug targets substantially increases the likelihood of a drug being approved for clinical use – up to two-fold!1,2 The hope is that this can help counter the trend dubbed Eroom’s law: the number of new drugs passing clinical trials for every $1 billion invested has been decreasing over time since the 1950s.3

The festival was full of examples of genomics being used for drug discovery. A wonderful talk by Paul Nioi of Alnylam Pharmaceuticals explained how his company’s burden tests of whole exome sequence uncovered a novel gene association for type 2 diabetes.4 On a panel, Charles Paulding of Regeneron explained how PheWAS of drug target genes could be used to identify potential adverse effects of drugs. Neil Ward of Illumina presented the fascinating story of sclerostin protein, produced from the SOST gene. Loss of function mutations in SOST lead to rare diseases of excessive bone mass, identifying it as a regulator of bone mineral density.5 Antibodies against sclerostin are now an approved treatment for osteoporosis, a common disease characterised by weak bones.

But it’s not all about drug discovery. The dream for many is personalised medicine: where your unique characteristics (including your genome) and views are used to deliver you your ideal treatment plan. There is a long journey before we reach this utopia. For example, today many individuals suffer within rare genetic disorders for years without knowing the cause of their disease.

British biotech company Congenica wants to change this. The company’s purpose is create software that translates next generation sequencing data into informative reports for clinical professionals. These reports can speed-up the diagnosis rare diseases caused by de novo mutations. Congenica works closely with the UK’s National Health Service (NHS) and has been a key partner during their 100,000 Genomes Project, a project which aims to sequence 100,000 whole genomes from NHS patients affected by rare diseases and cancer.  

Connecting Patients to Genomics

The 100,000 Genomes Project brings patients closer to their own genomic data as it directly influences their care. As well as this, the data can be used in research. The 100,000 Genomes Project makes abstracts of research being performed using patient data publically available to foster patient engagement. The All of Us research programme, which aims to build a biobank of 1 million Americans for research purposes, has pledged to do the same.

All of Us will also go one step further: returning non-health related insights from people’s genomes to them, if they so wish. This would include information about genetic ancestry and traits such as eye colour. All of Us state that many people they surveyed were interested in receiving this sort of information, which would commonly be purchased from a company such as 23&Me.

Most research cohorts do not return any health-related genetic data to participants. Doing so raises a lot of ethical and logistical hoops to jump through. Those that are most successful at doing so are the projects built in collaboration with healthcare providers, such as the NHS and the 100,000 Genomes Project. Another example is MyCode, a massive whole exome sequencing project launched by US private healthcare provider Geisinger. Patients of Geisinger can opt in to exome sequencing and share their medical records with the project. Geisinger are offering to return any clinically actionable results to any patient who agrees to receive them. They claim that these have reduced in incidence of preventable disease within their patients.

Patient involvement can directly influence research in a positive way beyond just providing data. Nick Sireau, chair and CEO of a patient advocate group for the rare genetic disease alkaptonuria (the AKU Society), gave an excellent presentation on how their work with pharmaceutical companies contributed to the approval of the first ever treatment for AKU. Listening to patients is a simple but often overlooked way to gain novel insight into a disease.

Connecting Researchers to Data

Another thing that comes along with the union between biobanks and healthcare providers is access to electronic health records. The UK Biobank’s (UKBB) linkage to NHS health records is a prime example. This provides researchers with longitudinal data that would otherwise require large quantities of time and grant money to generate.

However, this so called ‘real world’ data has not been collected for the purpose of being studied by academics. This leaves the door open for unseen biases to creep in. For example, differences in the way that groups of individuals interact with their healthcare system. In addition, there are usually data quality issues that need to be addressed before real world data can be used in an analysis.

Nevertheless, read world data is a rich data set that can be called upon as another piece of supporting evidence for a hypothesis. Ideal for say, proposing the repurposing of an already in-use drug. With appropriate statistical modelling and careful interpretation, there are bound to be golden discoveries hidden in this messy data treasure chest.   

A prominent issue in genomics that may also creep into real world data is the poor representation of minority populations. It is absolutely imperative to increase the diversity of our genomic datasets. Without this, we cannot give the same standard of care to minorities as we are to others. We are also doing our research a complete disservice. Underrepresented populations may carry unique genetic variation that could help us better understand complex traits. Ideally, individuals from underrepresented groups should be at the core of the collection and analysis of these datasets to prevent exploitation.

This will likely require that genomic datasets and the compute resources to analyse them become more accessible. Both UKBB and All of Us announced that they would be releasing web-based data platforms in the near future. Sequence data and deep phenotyping mean that huge amounts of storage space are required for all of the data associated with a large cohort. It is therefore more sensible to give researchers access to a portal where they can access the data, rather than each group storing the data individually. UKBB’s set up will have integration with cloud provider AWS, allowing researchers to carry out analyses on the data within the platform, paying only for resources used. All of Us also plan for researchers to be able to perform analyses within their web platform and will be providing integration with commonly used tools such as Jupyter Notebook. This will make performing genomics research more accessible to researchers who do not have their own high performance computing infrastructure.

Such platforms will allow large cohorts to continue growing their datasets. Both Genomics England and UKBB expressed their desire to expand their –omics beyond the genome. UKBB will be releasing metabolomics assay data later this year and have plans for proteomics assays. These –omics data will require even more storage, making the data platform approach a necessary tool of the future.

Final Comments

The festival really hit home that we are accelerating towards a future of medicine shaped by human data. I, for one, find this extremely exciting. In fact, it is the reason that I started my PhD. But perhaps more importantly, it really outlined what was necessary to achieve that future: connection. Connection between academic researchers, patients, healthcare providers, and biotech companies. Genomics is a group effort: attempts to go it alone are futile. Everyone has something to learn from everyone else. This inter-dependence is something beautiful: it promotes genomics for good.

References

1.         King EA, Wade Davis J, Degner JF. Are drug targets with genetic support twice as likely to be approved? Revised estimates of the impact of genetic support for drug mechanisms on the probability of drug approval. PLoS Genet. 2019;15(12):e1008489. doi:10.1371/journal.pgen.1008489

2.         Nelson MR, Tipney H, Painter JL, et al. The support of human genetic evidence for approved drug indications. Nat Genet. 2015;47(8):856-860. doi:10.1038/ng.3314

3.         Scannell JW, Blanckley A, Boldon H, Warrington B. Diagnosing the decline in pharmaceutical R&D efficiency. Nat Rev Drug Discov. 2012;11(3):191-200. doi:10.1038/nrd3681

4.         Deaton AM, Parker MM, Ward LD, et al. Gene-level analysis of rare variants in 363,977 whole exome sequences reveals an association of GIGYF1 loss of function with diabetes. medRxiv. January 2021:2021.01.19.21250105. doi:10.1101/2021.01.19.21250105

5.         Lewiecki EM. Role of sclerostin in bone and cartilage and its potential as a therapeutic target in bone diseases. Ther Adv Musculoskelet Dis. 2014;6(2):48-57. doi:10.1177/1759720X13510479