Teamwork is crucial for proteins. Little is known, however, about which protein teams are actually active in which tissues. A new large-scale study by systems biologists at ETH Zurich is now redrawing the map.

In brief
- Researchers at ETH Zurich mapped protein interactions in eleven human tissues using over 7,800 human tissue samples.
- Around a quarter of the protein interactions the researchers found are limited to a single tissue.
- The findings could improve the search for new drugs with fewer side effects.
The human body and its organs are composed of a wide variety of cell types. Although all cells contain the same genes, they function very differently - partly because protein interactions differ between cells.
Researchers at ETH Zurich have now created an atlas showing which proteins work together in which tissues. These findings will help to identify disease genes more precisely and develop drugs that act specifically where they are needed and nowhere else.
Proteins seldom work alone, but in teams: they interact with other proteins, for example by forming complexes or along biochemical signalling cascades.
In order to better understand how cells function in different tissues or organs of the body, scientists must first find out which proteins are working together and how their coordination differs from cell type to cell type. "If we know the specific protein interactions, we can better understand what distinguishes a liver cell from a brain cell," says Pedro Beltrao, professor at the Institute of Molecular Systems Biology at ETH Zurich.
Sifting through huge data sets
However, it is not yet sufficiently established which of these protein interactions occur universally throughout the body or are specific to a particular tissue. One reason for this is that systematically investigating protein interactions by way of laboratory experiments is expensive, time-consuming, and complicated.
Consequently, Beltrao and his colleagues opted for bioinformatics and tapped existing data collections on the proteome (the totality of all proteins present in a cell at a given time).
In their study, which has just been published in the journal Nature Biotechnology, the ETH researchers combed through proteome data from over 7,800 human biopsies. Based on these huge data sets, the researchers then inferred tissue-specific protein interactions in eleven different tissue types.
Every fourth interaction is tissue-specific
The ETH researchers found that every fourth protein interaction is indeed tissue-specific. For example, one interaction may occur only in liver tissue, but not in the other ten tissue types examined. Beltrao attributes most of the differences between tissues to cellular components that are specific to cell-types. Cellular compartments are distinct areas within a cell where certain activities take place - which is where some protein interactions occur.
A particularly high number of tissue-specific protein interactions occur in the brain and its neural connections (synapses). In order to examine the protein teams in synapses, Beltrao and his colleagues used data collected by ETH Professor Bernd Wollscheid in his laboratory.
Better drugs targeting specific areas
The findings are also valuable for pharmaceutical research. Knowledge of organ-specific protein interactions helps researchers to better understand disease mechanisms and identify disease genes to improve drug development.
„Proteins engaging in teamwork usually also influence the same disease," Beltrao explains. "So if we can say which proteins work together exclusively in nerve cells, for example, we can better define the genes involved in the disease."
Today's drugs can have a broad spectrum of activity and, as a result, incur undesirable side effects in parts of the body beyond the therapeutic target. The findings from this study may help researchers to search for active substances that act on proteins in a specific organ or tissue. This will increase drug safety thanks to specificity.
Reference
Laman Trip DS, van Oostrum M, Memon D et al. A tissue-specific atlas of protein-protein associations enables prioritization of candidate disease genes. Nature Biotechnology (2025). external page doi: 10.1038/s41587-025-02659-z