Diseased cells, such as cancer cells, often show clear differences in the type and quantity of proteins present. However, while diagnostics increasingly consider genomic alterations, this has only been possible to a limited extent with the more complex protein analysis. Technical progress in this field opens up new opportunities, but only in combination with new, suitable evaluation methods: The reliability and user-friendliness of the various methods can differ greatly, especially when examining real patient samples. Researchers at the University Medical Centre Freiburg and the Cluster of Excellence CIBSS – Centre for Integrative Biological Signalling Studies at the University of Freiburg now found combinations of methods that are particularly good at detecting the altered protein patterns of diseased cells. The study was published in the journal Nature Communications on 12 May 2022.

For their analysis, the team led by CIBSS member Dr. Clemens Kreutz and Prof. Dr. Oliver Schilling compared almost three million evaluations. The researchers also showed that certain open source programmes are well suited for analysing the complex data. Their results are an important step towards the future establishment of protein analysis as a diagnostic tool in clinical application and therapy.

"Our study shows very clearly that protein analysis can be a good method to further refine diagnostics, for example in cancer. But this form of diagnostics could also become relevant in the future for many other medical conditions such as cardiovascular diseases," says Schilling, Heisenberg Professor for Translational Proteomics at the Institute of Clinical Pathology at the University Medical Centre Freiburg.

Basis for new diagnostic approaches

The researchers, including CIBSS member Eva Brombacher, examined lymph node tissue from 92 cancer patients. Using mass spectrometry, they determined the type and quantity of more than 7,000 proteins. They analysed the measurement results together with scientists from the Institute of Biometry and Statistics at the University Hospital in Freiburg and compared 1,400 available analysis programmes in around 2,100 different data sets. In the study, they now show which of the approximately 1,400 combinations are suitable for analysing human samples and which are not. "We have thus created an important basis for the development of new diagnostic methods," adds Kreutz, who is a scientist at the CIBSS and Institute for Medical Biometry and Statistics (IMBI) at the University Medical Centre Freiburg. In international research networks, they continue to investigate whether disease progression or response to therapy can be predicted on the basis of altered protein patterns.

Original publication:

Fröhlich, K., Brombacher, E., Fahrner, M. et al. Benchmarking of analysis strategies for data-independent acquisition proteomics using a large-scale dataset comprising inter-patient heterogeneity. Nat Commun13, 2622 (2022). https://doi.org/10.1038/s41467-022-30094-0