Microglial Activation ODE Model — TREM2/APOE/IL-6 Signaling Network

🧮 Model Details Ordinary Differential Equations (Ode)

Internalcandidate

ARCHITECTURE

scipy.integrate.solve_ivp

METRICS

parameter_count

training_config

{'equations': ['dM_homeostatic/dt = -k1*TREM2 + k2*IL6 + k3', 'dM_dAM/dt = k1*TREM2 - k4*M_dAM - k5*M_dAM*Abeta', 'dTREM2/dt = k6*APOE - k7*TREM2 - k8*TREM2*amyloid_load', 'dIL6/dt = k9*M_dAM - k10*IL6'], 'data_source': 'SEA-AD bulk RNAseq + published kinetics', 'n_parameters': 12, 'fitting_method': 'least_squares', 'state_variables': ['M_homeostatic', 'M_dAM', 'TREM2_surface', 'IL6', 'phagocytosis_rate']}

evaluation_metrics

{'AIC': -142.3, 'RMSE': 0.023, 'R_squared': 0.94, 'n_data_points': 84}

EVALUATION CONTEXT

Dataset:In silico — ODE fit against SEA-AD bulk RNAseq + published kinetics (84 data points); R²=0.94, RMSE=0.023, AIC=-142.3

Benchmark:TREM2/APOE/IL-6 Signaling ODE Model — Microglial Activation Benchmark

Training data:SEA-AD bulk RNA-seq expression data + published TREM2/APOE/IL-6 neuroinflammation kinetics (Gomez-Nicola & Perry 2015; Heneka et al. 2015)