Why Synthace is collaborating with Microsoft Research to change the way research on programming biological systems is conducted
Despite great advances in our understanding of fundamental biology, ways of conducting research have advanced little since the 1960s. Experiments are still often conducted manually using artisanal techniques, with results recorded in paper notebooks or typed up by hand into electronic lab notebooks (ELNs). Even though there already exists a plethora of expensive lab equipment, too often this sits idly in store rooms or, if in use, isn’t utilised to its full potential.
If labs have had access to the hardware for some time, why is modern biological research still overwhelmingly artisanal?
At Synthace, to answer these questions we talked to our customers. What became clear was some of the barriers that have prevented widespread adoption of automation included:
1. Poor user interfaces on devices (often requiring a significant time investment to set-up experiments)
2. A lack of connectivity/standards/commonality of UI between different pieces of equipment
3. An over focus on inflexible automation solutions for high throughput screening
And yet, at the higher levels of the biopharmaceutical industry, there is a renewed focus on improving productivity through technology – one of these being automation, another being machine learning.
But why will it be any different this time? Why will automation finally deliver the productivity gains it previously promised?
Computer Aided Biology: The Future
The answer is: a new generation of software will enable new ways of working with biology.
At Synthace, we call this new way of working – powered by software - “Computer Aided Biology” or CAB. CAB is to Engineering Biology, what Electronic Design Automation was to electronics engineering; or what CAD/CAE/CAM was to automotive engineering.
It is the integrated ecosystem of software tools that enable scientists to quickly work in an iterative data driven way to ‘program’ systems of increasing biological complexity.
We are already seeing some tentative steps towards this way of working, exemplified by the closed loop of design, build, test and analyse. However, in nearly all cases this focuses on low complexity/high volume serendipitous approaches such as Hit/Led generation (e.g. screening), with the bottleneck then being moved further down the R&D pipeline. In addition, these strategies, whilst useful in scenarios where existing compound libraries are screened (e.g. small molecules), they are less useful in engineering or programming bespoke ‘living’ therapies (e.g. a CAR-T Cell Therapy).
To see the benefits of ‘automation’, more tools are needed that change the way of working across the whole pipeline of R&D; tools that allow us to apply engineering principles to rationally design, simulate, build and test new systems. However, when new ‘design’ software tools create an increased volume of rationally programmed biological designs for experimental validation, and machine learning tools can rapidly analyse results from said experiments - this necessitates a new breed of digital to physical translation engines, such that the design, build, test cycle isn’t limited by manual, expensive and error prone wet lab experimentation. This is where the future value of automation lies – integrated with digital tools, allowing the rapid translation of the digital into the physical, before churning our data back for digital tools.
Thanks to the emergence of infrastructure providers, such as Microsoft Azure; a growing number of fast moving and innovative start-ups are now building ‘best in breed’ products for biological research; without prohibitively expensive infrastructure overheads or having to invest in an in-house security team. This has resulted in an explosion in machine learning tools for programming biology (Bio-CAD), Data Management and Analysis packages; whilst a limited number of providers such as Synthace have focussed on the digital to physical translation of bio designs into structured experimentally validated data sets.
By collaborating with Microsoft Research, Synthace hopes that we can employ the expertise Microsoft brings in machine learning, cloud infrastructure and network driven systems biology to accelerate the production of new biological materials and therapeutics. For more details on the collaboration, check out the blog on the Microsoft Innovations page.
So where next for Computer Aided Biology?
A tsunami of data, not just from the biological laboratories, but from real world sources such as sensors and patient records, means we finally have a hope of getting the right treatment, to the right stratified group of patients at the right time – something often called precision medicine. But to enable this, comprehensive interoperable data standard across the research, development and care pathways need to be established; along with automated ‘smart’ AI systems for data collation, experimentation and analysis. In this future vision, Computer Aided Biology and the ability to systematically program biology, linked to organisational wide cloud-based data lakes, will be the key to enabling a fast-moving pipeline of patient specific, safe & cost-effective new therapies.