Table 3 Accuracy of direct and meta-prediction of MDS-UPDRS I, II, III, and total progression.

From: Genetically-informed prediction of short-term Parkinson’s disease progression

 

PPMI train-test split

PDBP independent test

Direct prediction

12-month MDS-UPDRS I

0.70 ± 0.04

0.70 ± 0.05

12-month MDS-UPDRS II

0.69 ± 0.04

0.70 ± 0.05

12-month MDS-UPDRS III

0.74 ± 0.04

0.73 ± 0.05

12-month MDS-UPDRS Total

0.74 ± 0.04

0.71 ± 0.05

24-month MDS-UPDRS I

0.70 ± 0.04

0.69 ± 0.05

24-month MDS-UPDRS II

0.73 ± 0.04

0.72 ± 0.05

24-month MDS-UPDRS III

0.76 ± 0.04

0.74 ± 0.05

24-month MDS-UPDRS Total

0.75 ± 0.04

0.73 ± 0.05

36-month MDS-UPDRS I

0.76 ± 0.04

0.75 ± 0.05

36-month MDS-UPDRS II

0.72 ± 0.04

0.74 ± 0.05

36-month MDS-UPDRS III

0.74 ± 0.04

0.77 ± 0.04

36-month MDS-UPDRS Total

0.67 ± 0.04

0.74 ± 0.05

Meta-prediction

12-month MDS-UPDRS Total

0.77 ± 0.04

0.75 ± 0.05

24-month MDS-UPDRS Total

0.76 ± 0.04

0.75 ± 0.05

36-month MDS-UPDRS Total

0.77 ± 0.04

0.73 ± 0.05

  1. F-measure accuracy of direct (single model) MDS-UPDRS I, II, III, and Total progression prediction (top) and F-measure accuracy of MDS-UPDRS Total meta-prediction (bottom). PPMI is used for training in all cases. Train-test split is 75% training and 25% testing with stratified sampling. See Fig. 2 for feature importance for MDS-UPDRS I, II, and III direct prediction and Fig. 3 for feature importance for MDS-UPDRS Total meta-prediction. ROC and PR curves for meta-prediction are provided in Fig. 4.
Back to article page