The power of many: Combining published datasets to identify Alzheimer’s disease biomarkers in cerebrospinal fluid

ABSTRACT: This manuscript describes methodology for the meta-analysis of published proteomics datasets from multiple publications to mine for cerebrospinal fluid (CSF)-based diagnostic biomarkers for Alzheimer’s disease (AD).The identified biomarkers are then validated in an independent cohort analyzed in house and two additional published large datasets. Specifically, the published data was mined from 6 independent and a 7th in-house acquired CSF proteomic datasets resulting in a meta-cohort with 73 AD cases and 77 controls. In-depth data analysis revealed 35 CSF biomarker candidates, many of which are associated with brain glucose homeostasis, which is described to be dysregulated in AD. Next, the list of identified biomarker candidates were then validated in a 2-pronged approach using i) an independent cohort analyzed in house, and ii) two recently published independent large scale datasets comprising in total more than 500 samples. The resulting biomarker panel consisting of three glycolytic enzymes was found to discriminate AD from controls. To our knowledge, this is the first study that described the mining and systematic re-analysis of previously published liquid chromatography mass spectrometry-based proteomics data to identify and validate biomarkers in general and CSF biomarkers for AD in particular. While the presented study focused on AD, the presented workflow can be applied to any disease for which published datasets are available.

INSTRUMENT(S): Q Exactive

ORGANISM(S): Homo Sapiens (human)

TISSUE(S): Cerebrospinal Fluid

DISEASE(S): Alzheimer's Disease

SUBMITTER: Patrick van Zalm  

LAB HEAD: Dr. Hanno Steen

PROVIDER: PXD022649 | Pride | 2023-02-16

REPOSITORIES: Pride