Extended Data Fig. 2: Prediction of incident dementia using proteins, demographic and clinical variables, and their combination measured at late-life baseline (2011-2013).
From: Large-scale plasma proteomic analysis identifies proteins and pathways associated with dementia risk
Elastic net machine learning with Cox proportional hazards regression was used to select the best combination of proteins from the top 50 proteins in each model. This table shows results from 10-fold cross validated analyses. Two-sided P-values were calculated for the C statistic comparisons. No corrections for multiple comparisons were performed. a Protein combination defined using elastic net machine learning algorithm. b Includes age, sex, race-center, education, and APOEε4. c Includes body mass index, diabetes, hypertension, smoking status, and eGFR-creatinine. Abbreviations: C stat. Δ, change in C statistic with the addition of elastic net proteins.
