Robots to Replace Scientists

 

By Susan Sheng

[quote style=”boxed” ]Life science research today is incredibly slow, error-prone, monotonous, and expensive with researchers spending many hours a day every day just moving small volumes of liquids from one place to another.[/quote]

Transcriptic

 

If I had to make a wish-list of things that would be helpful in my daily life, a robot to help me with the more repetitive and mundane lab tasks would be at the top of that list. Whether it’s pipetting samples to run a Bradford assay or running a western blot to look at changes in protein expression, there are many tasks in a typical biological sciences laboratory that are tedious and time-consuming, but necessary for answering scientific questions. While certain techniques, such as DNA sequencing, have benefited from advances in technology and automation, for many other techniques, either an automated option does not exist, or the machines necessary to run the protocols are too expensive for most research labs to consider. Luckily, as highlighted in a recent Nature commentary, there is growing interest in automating research in the hopes of improving its efficiency and reproducibility.

 

Two California companies, Emerald Cloud Laboratories and Transcriptic, are trying to offer cost-effective access to wet-lab experiments. Both offer web interfaces which allow scientists to design wet-lab experiments which are then carried out at a remote facility. Emerald Cloud Laboratories, which started as an internal system to streamline research at the biotechnology company Emerald Therapeutics, offers 40 common lab protocols, including western blotting, light and epifluorescence microscopy, and DNA/RNA extraction and downstream procedures such as gel electrophoresis and PCR, with plans to add many more procedures to the list. Similarly, Transcriptic allows users to program experiments using their application programming interface which are then carried out in their facility. They aim to transform basic research and lower the entry costs such that anyone with an idea and web access can start a biotech company, much like how the computer science world currently operates.

 

If companies like Emerald Cloud Laboratories and Transcriptic succeed in offering reliable, low-cost research services, it could revolutionize the way research is done. By outsourcing basic, repetitive tasks, scientists can be freed up to plan and design the next experiment, or carry out more technically challenging experiments. Additionally the increase in automation could help address the issues of reproducibility in science, which have been discussed here at the Scizzle blog  among other places. Automation would bring another level of standardization and documentation to experiment design and methods, and remove problems of human variability and error. I’m sure we all have received protocols that have worked wonderfully for other people, only to struggle to successfully repeat the protocol in ourselves. Of course, automated experiments will still need to be optimized and troubleshot for each specific condition or context, but at least the factor of human error is reduced. Additionally, automation could improve the efficiency and speed in which research is done; for instance, the Allen Institute for Brain Science was able to complete in situ hybridizations in adult mouse brain tissue for over 20,000 genes in a matter of weeks due to the automation of several steps in the process.

 

It’s unlikely that robots could replace research scientists, at least for the foreseeable future. We need trained scientists to come up with the questions and the experiments that will advance our knowledge. Tasks requiring a high level of precision and dexterity will likely still require human hands, at least until that level of precision can be achieved robotically. However, robots working alongside researchers could speed up the scientific discovery process. Whether automated processes take hold in the life sciences will likely depend on cost, reliability, and the ease it which standard protocols can be customized for different experimental conditions. I for one am excited to see what the future of research has in store.