Model does not replace human researchers but rather enhances their capabilities, say authors
Photo: iStock A new metabolic protein that can tolerate higher temperatures have been developed, not by humans but an autonomous robotics lab run by artificial intelligence!
The platform called self-driving autonomous machines for protein landscape exploration (SAMPLE) was designed by scientists from the biochemistry department of the University of Wisconsin–Madison, United States.
Protein engineering is a field with vast potential, touching diverse areas such as medicine, energy and chemistry. However, traditional methods of creating proteins with new or improved functions are notoriously slow, demanding and inefficient. The process typically involves generating hypotheses, designing and conducting experiments, and interpreting results — a cycle that is both labour-intensive and time-consuming.
The SAMPLE platform is a leap forward in this context. It operates as a fully autonomous system, leveraging artificial intelligence capable of learning protein sequence-function relationships, according to the report published in the journal Nature Chemical Engineering.
The researchers designed SAMPLE to do the job of a protein engineer but without needing people to run the experiments.
Once it learnt how a protein’s structure impacts what it does, the model used that knowledge to create new protein designs, the authors wrote. (Think of it like a chef learning what ingredients taste good together and then coming up with a new recipe.)
SAMPLE didn’t stop at designing these proteins on paper: The system then instructed robotic instruments (lab assistants, if you may!) to put together the new protein based on the design and then test it to see if it works as expected, according to the scientists who wrote the paper.
The feedback from these tests enhanced its understanding of protein structures, creating a continuous cycle of learning and discovery, the authors added.
In their experiment, they focused on a type of protein called glycoside hydrolase. They wanted SAMPLE to make versions of this protein that could handle higher temperatures.
After 20 rounds of experimentation in over six months, the self-driving lab produced new enzyme versions that could function at temperatures at least 12 degrees Celsius higher than the starting proteins. This is important because proteins that can withstand heat can be very useful in industries like biofuels and medicine.
Done manually by a human, the work would take about a year, the authors noted. It's an example of how AI and robotics can take over repetitive scientific work, allowing scientists to focus on more creative and complex tasks, they stressed.
SAMPLE's implications are far-reaching. It promises to significantly speed up the process of scientific discovery in protein engineering and synthetic biology. Moreover, its ability to operate without human intervention reduces the chances of error and improves the efficiency of the protein engineering process.
While SAMPLE represents a significant advancement, challenges remain. Biological experiments are inherently complex, involving high-dimensional genomic spaces and nonlinear phenotypes. Furthermore, integrating advanced analytical instruments for broader applications remains an area for further development.
Moreover, SAMPLE's approach to learning and optimisation could pave the way for the discovery of proteins with functions that are currently unknown or considered too complex to engineer using traditional methods.
The role of human researchers remains critical. The design of initial hypotheses, selection of target functions and interpretation of broader implications of the findings still require human expertise and insight. Thus, SAMPLE does not replace human researchers but rather enhances their capabilities, allowing them to focus on more creative and complex aspects of protein engineering.
We are a voice to you; you have been a support to us. Together we build journalism that is independent, credible and fearless. You can further help us by making a donation. This will mean a lot for our ability to bring you news, perspectives and analysis from the ground so that we can make change together.
.png)
Comments are moderated and will be published only after the site moderator’s approval. Please use a genuine email ID and provide your name. Selected comments may also be used in the ‘Letters’ section of the Down To Earth print edition.