FIGURE 1.

A graphical overview depicting development of the CSF aging clock, model feature extraction, and bioinformatics. (1) CSF from 437 cognitively normal, amyloid beta negative individuals were measured using the SomaLogic system to quantify 7008 proteins. (2) An elastic net was trained on the proteomics profiles to develop a machine learning algorithm that predicts biological CSF age. (3) Proteins used to predict age were extracted from the model, ranked and categorized into older associated and younger associated groups. (4) Pathway enrichment was performed via the bioinformatics tool Metascape to look for pathways enriched in model proteins in the Reactome pathway knowledgebase.