demos

Goldilocks DPM Demos

Demonstrations of the Goldilocks-DPM framework for data-driven Disease Progression Model configuration.

Contents

1. goldilocks-pysustain.py: Subtype and Stage Inference (Young et al., Nature Communications 2018)
2. Future feature (TODO): goldilocks-ebm.py: Event-Based Model (Fonteijn, et al., NeuroImage 2012)

Workflow for Z-Score SuStaIn

See goldilocks-pysustain.py for a worked example, but here's a conceptual overview.

• Import the module:

from goldilocks_dpm import goldilocks_ZscoreSustain

• Prepare your data.
  • SuStaIn: see examples in pySuStaIn for inspiration and guidance.
  • EBM: see kde_ebm and disease-progression-modelling.github.io for inspiration and guidance.

• Create a goldilocks ZScoreSustaIn object including your data matrix X and vector y labelling cases and controls (controls are used to z-score your data):

output_folder = Path.cwd() # or wherever you want the output to go

gdpm = goldilocks_ZscoreSustain(
    dpmData = X,
    classes = y,
    output_folder = output_folder,
    robust_zscores = False,
    case_label = 1,
    ctrl_label = 0, 
    direction_abnormal = direction_abnormal,
    biomarker_labels = biomarkers
)

• Run the goldilocks zone calculation

gdpm.run_goldilocks(
    plot = True,
    plot_format = "png",
    verbose = False
)

• Your goldilocks_ZscoreSustain object will now contain the z-scored X data, and goldilocks-suggested pySuStaIn hyperparameters Z_vals and Z_max