Perkins says the opportunity for humanity -- and Human Longevity -- is the result of the convergence of four trends: the reduction in the cost of genome sequencing (from $100m per genome in 2000, to just over $1,000 in 2014), the vast improvement in computational power, the development of large-scale machine learning techniques and the wider movement of health care systems towards ‘value-based’ models. Together these trends are making it easier than ever to analyse human genomes at scale.
As CMO, Perkins leads all clinical and therapeutic work at Human Longevity Inc. Previously he was an executive vice president at Vanguard Health Systems, and prior to that worked at the US Center for Disease Control, where he led the 2001 investigation into anthrax attacks. Now at Human Longevity, he believes the techniques used to analyse diseases in the previous century can now be applied to the root causes of human aging.
The "highly secret" Human Longevity business plan is actually very simple, Perkins jokes: the company wants to take large amounts of genome data, subject it to machine learning and analytics, and "radically, disruptively" produce new models for medical care. "Our focus is not being a fee for service sequencing operation," Perkins says. It is to "fully understand and fully interpret all the meaning in the human genome". To do that Human Longevity Inc is building machine learning systems which can act as a ‘Google Translate’ for genomics, taking in genetic code and spitting out insights.
To make that possible they’ve hired as their chief data scientist Franz Och, who was formerly at Google working, among other things... on Google Translate. "To translate the language of biology in the form of sequence data into the language of health and disease, into the form of clinical phenotypes" is the aim, Perkins said. "To define and continuously evolve what we consider to be a 21st digital description of the full human phenotype."
The announcement of the first fully-sequenced human genome was made in 2000, but since then some have begun to lose faith that the area will produce genuine health insights, Perkins admits. But he sees that as a natural stage in the progression of the technology, and says we are now on the verge of the first transformative combination of analytics, data and health science.
To help move that revolution forward Human Longevity Inc will have to analyse the whole genome sequence -- that means 3.2bn base pairs, compared to much lower numbers analysed by existing, mainstream DNA research companies such as 23andMe. To collect and analyse all that data requires the world’s largest sequencing centre, which Human Longevity Inc operates, and estimates is currently able to sequence 35,000 human genomes per year. That will almost triple to around 100,000 genomes per year by the end of 2015 and target is 1 million genomes per year by 2020. "I’d like to see that number higher," Perkins says. Whatever the technology is needed to make that happen -- whether it’s an evolution of the Illumina HiSeq X Ten and Pacific Biosciences platforms Human Longevity already uses, or something else entirely, Perkins is open to change.
The results, he believes, will be revolutionary -- and make genuine differences in people's lives -- including his own. "My daughter is graduating from university next month, my father if he were alive would be 78 years of age... I’m encouraged that we’re on the verge of having lots more grandfathers and grandmothers at the special events of all of our lives," Perkins says. "As genomics begins the process of revolutionising human health and the practice of medicine, and opens the door to the next steps... of regenerative medicine. It’s going to be an extraordinarily exciting ride."
This article was originally published by WIRED UK
