Stripping the Complexity from Complex Experimentation

Biology is the most complex of all scientific fields. The living world has been expertly fine-tuning itself through evolution for circa 3.5 billion years, providing us with a myriad of organisms and interacting systems. Whilst we have been studying biology for hundreds of years, we still lack a true understanding of all the complexities of biological systems, meaning that we are yet to harness the full potential of biology.

An engineering approach is rigorous and reproducible. We’re getting better at addressing that challenge in biology - but there’s still a long way to go!

The rate of biological discovery is escalating, providing scientists with the molecular and technological tools necessary to engineer biological systems for predefined outcomes – to produce medicines, chemical commodities, novel materials and more. Yet due to the interlinking intricacies of biology, we still have some way to go before we can truly call it an engineering discipline, with all the rigour and exacting precision that the latter entails So, how can we drive forward our understanding of the biological systems we are trying so hard to engineer?

Forget what you learned in academia – changing one factor at a time shows only a fraction of the real picture

The past (and still current) biological practice of varying one experimental parameter at a time and measuring a change in output, just to move onto the next parameter, is rapidly becoming a redundant methodology; practices like this will never truly reveal the depth and breadth of knowledge required about a biological system for us to robustly engineer it. Understanding how one parameter varies in response to another parameter starts to provide finer details into the inner workings of biology, but still isn’t sufficient.

A more modern and robust approach to experimental planning and implementation is beginning to be adopted more widely, that of Design of Experiments (DOE). Using statistical methods to drive the design of experiments, with DOE we can now investigate upwards of 10 parameters at a time, with each of those spanning multiple levels (such as temperature, concentration or chemical additive).

DOE is demanding, but it’s increasingly accessible with automated liquid handling

DOE experimental procedures require between hundreds and thousands of individually unique experiments to be run to ascertain which of the investigated parameters influence the studied system and which of these parameters are interacting with one another, either positively or negatively. This style of experimentation starts to unlock the biological knowledge that we have been missing for years with the antiquated methodology.

It is feasible to run complex DOE designs manually, but this places the burden on the scientist of performing all the manual liquid handling – setting up potentially 1000 reactions – with the unrealistic notion of maintaining complete accuracy. It’s not just the time taken to set up the experiments that is an issue but also the hours to weeks of planning in preparation of such high-dimensional and complex experimentation. As we start to reach the boundaries of what is feasible in the manual implementation of this multifactorial methodology, we need to find new solutions to allow us to push the envelope of what is possible.

Software is needed to make DOE liquid handling experiments quick and easy to program and edit, as often as you need – bringing biology into the digital era

The advent of low-cost and accessible automated liquid handling platforms provides a basis for implementation but not necessarily the whole solution. These platforms are extremely good at what they do, but often their software isn’t designed for complex experimentation such as a DOE. That leads to many hours or days of programming to implement such high complexity experimentation, which raises the barrier of usage for many laboratory scientists. The software programming problem is not just a one-off burden either but a repeated one, with every different DOE requiring re-programming of the liquid handling platform.

At Synthace, we have created software, which affords its users simple, rapid and flexible re-programming of automated liquid handling platforms as well as integration of all lab hardware including analytics.

The Synthace software platform is compatible with commonly used DOE software packages such as JMP and Design Expert, parsing design files and translating them into the low-level liquid handling instructions required by liquid handling platforms. Instantly, the days to weeks of experimental planning and programming is shortened to minutes, safe in the knowledge that the DOE itself will be reproducible and accurate.

With Synthace, it is now possible – and easy for the user – to implement experimentations of ever-increasing complexity. We can now start to change the way we think about biological experimentation, being able to investigate upwards of 20 factors simultaneously across multiple levels, truly opening the door to understanding biology at the smallest detail. Bringing biology into the digital era will revolutionise our understanding, edging us closer to the realisation of truly predictive engineering of biology.