TL:DR Genetics is a software problem, and there’s plenty of problems to be solved
Today I’m happy to announce we are leading an investment of $30m in SOPHiA Genetics. SOPHiA has developed a platform used by over 300 hospitals and 1000s of clinicians globally, to help diagnose patients illnesses using genomic data.
The company is at the epicentre of radical changes in the way that both the healthcare industry and the pharmaceutical industry will operate in the future, driven by precision medicine — medical care designed to precisely optimise each treatment to suit individual patients. This is a radical change to how medicine has been done for the last 4,000 years and a revolution that is just beginning.
We are investing along with Invoke Capital and Alyclo, and new investors 360’ Capital, which brings the total raised by SOPHiA to $60M since it’s founding in 2012.
Why precision medicine?
Precision medicine, which uses genomic data as its cornerstone, is a paradigm shift in the way that the health and pharmaceutical industries will operate. Genomic data, the billions of lines of biological code that defines a huge amount of how we and our diseases work, will soon form the foundation of the decisions that lead to the diagnosis, choice of drug and perhaps eventually even be the vector for treatment to tackle many major and minor illnesses.
We aren’t alone in believing this. In the last year investment into companies using genomic data to diagnose hereditary diseases and cancers has exploded. By the end of Q2 2017, investors had already put $2B to work in the space and that was before recent announcements of $200M+ investments into consumer genetics company 23andMe and Chinese genomic data company Wuxi NextCode.
This new wave of investment has come off the back of government initiatives prioritising the use of genomic data in healthcare, from the NHS’s 100,000 genes challenge, to President Obama’s 1M Genome ‘Moonshot’ Project, to the Chinese Government’s commitment of $9B to research in precision medicine.
These programmes are all predicated on the belief that the future of precision medicine starts with your genome ( and that of your disease too ). While clinicians across the world already use genomic data to support their diagnosis of many diseases, we’re still just at the very beginning of the development of this new technology. The promise of identifying diseases far earlier, and developing personalised medicine to cheaply and quickly deal with them will not only improve and save uncountable numbers of lives but radically reduce the costs of our current health system. This won’t just improve healthcare, it will also massively reduce the cost as well.
But as Uncle Ben said, with great data, comes great challenges. Just as the explosion of data in other industries, from finance to social media, has required the development of a whole new range of technologies and infrastructure around it, the explosion of genetic data is going to require a whole new software stack to enable this revolution too.
But why now?
This isn’t the first time such grand statements about genomics have been made. However, in the last decade, the challenge of using genomic data for clinical diagnosis has moved from being primarily a chemical and hardware problem to a becoming a data problem. And data problems are solved with software solutions.
This shift towards being a data-driven industry has been driven by 30 years of change in the way genomic data is captured. The speed and cost at which genetic data can be sequenced has outpaced even Moore’s law for computing. And at the same time, the number of genetic diseases we can test for has increased fivefold.
What does this mean? We we can now test for more diseases than ever, and we can get data faster and cheaper than ever.
Whenever a rapid explosion of new data like this happens, there’s a whole new ecosystem of technologies that need to be built to take advantage of that data. We only have to look at how the personal computer and internet era fuelled the growth of companies providing servers, databases, virtual machines, operating systems and machine learning systems.. And it’s the emerging machine learning revolution that is the real answer to ‘why now?’ With the breakthroughs in the diagnostic and hardware market, software in the genetic industry would still be interesting. However it is the application of the latest machine learning techniques that means the software layer is a now huge opportunity to change the way global healthcare operates.
And this is where Sophia comes in. Sophia combines the best parts of a SaaS tool for difficult workflows, with the power of machine learning on huge datasets, with an active community of clinicians using the technology to improve outcomes.
This is a classic example of where machine learning can create a virtuous circle. First, build a powerful tool that is used by a community to collect verified data ( in this case, analysis of genetic variants ), then train your algorithms to make the output from that data set more accurate, which in turn means more people join the community, and submit more verified data.
This model has worked for Google in search, CityMapper in transportation, Amazon in logistics, and we hope Sophia in data-driven medicine. The beauty of this model is the virtuous circle created by Sophia when the platform is used.
In fact this model is even more important in data-driven medicine, where the rapid pace of change in the space means that new diagnostic kits and sequencers are bought to market so regularly that it is extremely difficult to ensure that the way any two kits or sequencer works guarantees a highly accurate result. SOPHiA solves this problem by developing algorithms that minimise the noise from different systems and verifies the outcome of a diagnostic based on processes used in each specific hospital.
As an example of the power of this platform, when a new diagnostic approach ( such as liquid biopsy) or a new sequencer machine comes to market after many years of development, Sophia can add these tools for clinicians to use on their platform in a single software sprint. So while it can take years, and millions of dollars, to develop a new diagnostic kit, SOPHiA can allow 1000s of clinicians to use the new kit with a single software deployment.
This is a realisation of the huge promise of software development applied to clinical genetics.
To achieve this vision, SOPHiA has bought together world leading talent in clinical genetics, bioinformatics and machine learning to build the product, and on top of that have managed to build a high performing software sales team to match.
Despite the many 1000s of companies we meet here at Balderton each year, and the GigaBytes of pitch decks and Excel models we sift through as we look at investments, my decision to make an investment comes down to three simple questions. Why that market? Why now? Why them? SOPHiA and the team Jurgi has built certainly meets that challenge.