You may be surprised to know our DNA Purley codes for proteins. These little molecules are responsible for all of us; from eye color, to organ activity. Scientists have studied the structure of proteins for decades now. Understanding their shape, function, and interaction is essential to grasping how the body works. Importantly, they also tell us how disease comes about when something goes awry.
The structure of two proteins combined together is called a complex. Complex formation is what drives chemical reactions in the body, though many of the interactions between them are still a mystery due to their complexity. Previously, scientists characterized these interactions through long-winded laboratory techniques, manually choosing which pairs to investigate.
Using AI to predict complex interactions
Research groups from UT Southwestern and the University of Washington have teamed up to make this process more efficient. Using AI and evolutionary data, they were able to predict 3D models of proteins and their interactions. The analysis was carried out in yeast, where the team looked for possible interplay between protein pairs. 699 interactions had previously been characterized in these microorganisms, but the computer program identified 806 new possible interactions. Although this output shows the method’s incredible and promising potential, the technology still does need to be refined.
Why is this technology so useful?
As proteins govern all interactions in the body, the insights this AI algorithm brings span far and wide, including providing information about bodily functions like cellular construction, metabolism, DNA repair, and cell maintenance. Also, this technology could be used to identify roles for proteins whose function is unknown. “The work described in our new paper sets the stage for similar studies of the human interactome and could eventually help in developing new treatments for human disease,” said Dr. Cong, one of the lead authors of the paper published in Science.
The information generated from this algorithm about the predicted complex structures is readily available to download online (https://modelarchive.org/doi/10.5452/ma-bak-cepc). This organized free access to the generated information will help progress the field of structural biology immensely.
Source study: Science – Computed structures of core eukaryotic protein complexes