Machine Learning Model Produces Maps of Related Proteins

A new machine learning tool developed by DTU Compute and the Department of Computer Science at the University of Copenhagen (DIKU) may help accelerate the development of proteins for broad applications in everyday life, including medicine, food, detergents, and plastics remediation.

The biotech industry is constantly searching for the “perfect mutation,” where properties from different proteins are synthetically combined to achieve a desired effect. In recent years, machine learning has been brought into biotech research to help form a picture of permitted mutations in proteins. The obstacle to this approach has been that different images are generated depending on the method used. Even if you train the same model several times, it can provide different answers about how the biology is related.

"In our work, we are looking at how to make this process more robust, and we are showing that you can extract significantly more biological information than you have previously been able to. This is an important step forward in order to be able to explore the mutation landscape in the hunt for proteins with special properties," says Postdoc Nicki Skafte Detlefsen from the Cognitive Systems section at DTU Compute.

A protein is a chain of amino acids, and a mutation occurs when just one of these amino acids in the chain is replaced with another. Since there are 20 natural amino acids, the number of mutations increases so quickly that it is completely impossible to study them all. Even with simple proteins, there are more possible mutations than there are atoms in the universe. It is not possible to test everything in an experimental manner, so you must be selective about which proteins you want to try to produce synthetically.

The DIKU-DTU Compute model works by drawing a map of proteins, making it possible to generate a candidate list of promising proteins for closer study.  By presenting the model for many examples of protein sequences, the model learns to place closely related proteins close to each other and distantly related proteins far from each other.

"[The model] enables us to talk about what a sensible distance target is between proteins that are closely related, and then we can suddenly measure it. In this way, we can draw a path through the map of the proteins that tells us which way we expect a protein to develop from to another, i.e., mutated, since they are all related to evolution. In this way, the [machine learning] model can measure a distance between the proteins and draw optimal paths between promising proteins," says Wouter Boomsma, Associate Professor in the section for Machine Learning at DIKU.

The researchers tested the model on data from numerous proteins found in nature, where their structure is known, and they can see that the distance between proteins starts to correspond to the evolutionary development of the proteins, so that proteins that are close to each other evolutionally are placed close to each other.

"We are now able to put two proteins on the map and draw the curve between them. On the path between the two proteins are possible proteins, which have closely related properties. This is no guarantee, but it provides an opportunity to have a hypothesis about which proteins it could be that the biotech industry ought to test when new proteins are designed," says Søren Hauberg, professor in the Cognitive Systems section at DTU Compute.

The unique collaboration between DTU Compute and DIKU was established through a new center for Machine Learning in Life Sciences (MLLS), which launched last year with the support of the Novo Nordisk Foundation. In the center, researchers in artificial intelligence from both universities are working together to solve the fundamental problems in machine learning driven by important issues within the field of biology.

The findings were published in Nature Communications.