For multivariate analysis, multiple logistic regression, fixed model and likelihood ratio method analyses were performed to ascertain the impact of different variables on the IPR-AASTRE-B and with the aim of adjusting for possible confounding effects.
For each data set, DS1, DS2, DS3, DS4, and DS5, there were three quantification methods investigated (1, 2, and L denote one-tissue compartment model, two-tissue compartment model, and likelihood estimation in graphical analysis, respectively).
This methodology implements a generalized least squares (GLS) model and likelihood ratio tests to resolve whether correlation between two characters is dependent on the underlying phylogeny [31], [32].
Tests for interactions were performed using Wald tests that include main effects and interaction terms in the model, and likelihood ratio testing, comparing the fit of the model with only main effects to the fit of the model with both main effects and interaction terms.
Outcome analysis will be performed using a Cox proportional hazards model and likelihood ratio test.
The model and likelihood based approach are implemented to re-analyze large SNP datasets of the Parkinson disease case and control samples collected from multiple human cohorts.
