Creating a ‘Smart Lab’: Optimizing Your Scientific Data
Complete the form below to unlock access to ALL audio articles.
BioBright, a company dedicated to developing a ‘smart lab of the future’ recently announced their collaboration with the Wellcome Trust Sanger Institute. A common challenge researchers face in today’s lab is the sheer volume of data generated during an experiment. BioBright aims to maximize the reproducibility of scientific data.
“The problem is we’re drowning in data…” explained Charles Fracchia, BioBright's CEO, in a recent press release.
Their tools are designed to enhance scientific protocols enabling scientists to more accurately collect, manage and analyze their data. Their voice recognition tool ‘Darwin’ and associated software have already been implemented by Bijan Pesaran of New York University. BioBright’s Darwin Sync software will now be introduced into the lab of Dr Gosia Trynka, Immune Genomics Group Leader at the Sanger Institute. The hope is that this software will help their current research — prioritizing gene candidates for drugs against auto-immune diseases.
We spoke to Charles Fracchia to learn more about BioBright, the solutions they offer, and to find out what their vision of a ‘smart lab’ looks like.
Laura Mason (LM): Could you tell me more about your professional background and BioBright?
Charles Fracchia (CF): I started as a biologist and bioengineer. I worked at IBM and as a subsequent graduate student at Harvard Med/MIT with the Department of Defense. Throughout my professional career, I was exposed to a lot of problems within laboratories both in bioengineering and academic settings.
Reproducibility and inadequate tooling were a common pattern and a need for a solution became apparent. We received a grant to start our own company from the Defense Advanced Research Projects Agency (DARPA) and I jumped out of my PhD program to become the CEO of BioBright.
LM: Can you tell me more about the reproducibility crisis? What is it about our current laboratory set up that is causing issues with reproducibility?
CF: Reproducibility is when you get an experiment from a colleague or literature and try to reproduce results. Often, you’re unable to. There is a multitude of limiting factors ranging from poor descriptions of methods, unknown factors of influence, to varied ability in materials. The fundamental problem is that biomedical research is extremely multidimensional, containing large numbers of variables that influence results at any given time. It is impossible to control all factors.
We’re not collecting all of the data that we could collect and that could be useful. Instrumentation settings are not routinely collected. Experimental variables are only collected when the scientist has an a-priori idea that a factor is essential. While estimates vary, studies published, estimate that the reproducibility crisis is wasting $28 billion per year in the US alone. In context, the US Pharma R&D expenditure in the US is around $50 billion.
LM: How can BioBright help optimize research?
CF: BioBright has built software that can help collect, manage, and analyze all of the data necessary to create an accurate picture of what happened in an experiment. We do that with what we call the Darwin Smart Laboratory platform. This platform has three major components.
Darwin Sync, which allows for auto synchronization of equipment data and extracts all of the metadata such as instrument settings and makes it available.
Darwin Speech, which is a voice assistant much like Siri or Alexa but tailored for biomedical applications. It is designed to understand complex scientific jargon used in the lab. This way scientists can take notes on the fly verbally as they’re doing their experiment. There is no longer a tension between doing an experiment and documenting an experiment.
Darwin Terminal is the equivalent of NASA mission control for the laboratory. It can help you make sense of the torrent data that pours through the system and helps you answer scientific questions faster and in a more natural way (using Darwin speech).
LM: How can scientists ensure they are recording and optimizing the data they generate?
CF: One of the core difficulties in scientific research is that you are at the edge of what is known. Good scientific research pushes into the unknown. It’s almost impossible to know a-priori what is going to be needed to validate or invalidate a hypothesis. Part of that has led to a mix-up between exploratory and confirmatory studies. One of the major sources of reproducibility errors. Exploratory studies are trying to make sense of what phenomena may be occuring. Confirmatory studies on the other hand should be designed to validate or invalidate hypotheses that were generated from those first exploratory studies. It’s nearly impossible for a scientist to know what to collect ahead of time.
BioBright inverts that thinking, making it possible to collect 100 or 1000 times more information than what was possible with prior technology and as a result, cuts down the time it takes to do both exploratory and confirmatory studies. Without corrupting the original data, scientists can leverage the BioBright platform to answer multiple hypotheses faster and in a data driven way. We make it possible for scientists to search the data and all the context it was collected in.
LM: Could you expand on the progress you have made since you presented your TEDX talk in December 2016?
TedX talk: Creating the smart laboratory tools of the future.
CF: We are receiving a lot of interest in the field. Some of the largest pharma companies and most prominent research institutes are using our platform to answer scientific questions they could not answer before. We were funded by DARPA to initially build this platform and we have demonstrated huge workflow improvements in neuroscience where we improved a critical workflow by more than 20x giving unprecedented precision to scientists using the platform. We are working with major vendors in the field to incorporate their data into our system and improve their existing products.
LM: What does a ‘smart lab’ look like to you? How do you see this ‘smart lab’ evolving over the next 10 years?
CF: First and foremost, BioBright has been pioneering this idea that the smart lab should augment our human capabilities as scientists. We believe the future of the smart lab is one that includes human intelligence know-how and intuition with the precision, accuracy and volume that machines can bring. A smart lab to me is a lab where human scientists can seamlessly ask scientific questions in a natural language and get quantitative answers from machines as well as qualitative information from their peers. With the kind of smart lab BioBright is building, scientists can explore the near infinite space of scientific possibilities with accuracy and creativity. In this lab, instruments auto report their info. Data is contextualized and available at all times.
I want to see a future where using the smart lab like ours can reduce a drug’s ‘time to market’ by a factor of ten or more. At the end of the day, what matters is to be able to discover and make these drugs available faster and more intelligently.
Charles Fracchia was speaking to Laura Elizabeth Mason, Science Writer for Technology Networks.