Robust Biomarkers for prediction of risk and disease

Important results from the project

The aim of the project is to assist largescale analysis of population-based biobanks to promote research on precision medicine. Five of the largest commercial platforms for largescale analysis of proteins and metabolites was evaluated in regards to coverage, overlap and handling of repeated analysis over time. The project has been summarized in a technical report that can be found in DiVA portal (http://umu.diva-portal.org/smash/record.jsf?pid=diva2%3A1713751&dswid=1444).

Expected long term effects

The project has revealed that the protein platforms are highly overlapping, same proteins are analysed, while the metabolic platforms are complementing. The quantitative methods are more stable than the semiquantitative, due to batch-effects. The exception seems to be Somalogic that report surprisingly stable semiquantitative values. The preferred strategy to handle batch-effects is to analyse samples in random order. An alternative is to use bridging-samples and systematically divide samples for analysis. The project increases knowledge on how to structure studies.

Approach and implementation

There is an increasing interest in Sweden and the world to turn the large sample biobanks into easily accessible and reusable data. We requested and sent the same set of biobank samples repeatedly to five of the largest analysis platforms for protein and metabolite analysis with the aim of evaluating how to best create these data. The result is useful for the planning of future largescale studies. The lead time to go from sample to data was long emphasizing the great value of robust databanks.