AbstractMotivationBiomedical research typically involves longitudinal study designs where samples from individuals are measured repeatedly over time and the goal is to identify risk factors (covariates) that are associated with an outcome value. General linear mixed effect models have become the standard workhorse for statistical analysis of data from longitudinal study designs. However, analysis of longitudinal data can be complicated for both practical and theoretical reasons, including difficulties in modelling, correlated outcome values, functional (time-varying) covariates, nonlinear effects, and model inference.ResultsWe present LonGP, an additive Gaussian process regression model for analysis of experimental data from longitudinal study designs. LonGP implements a flexible, non-parametric modelling framework that solves commonly faced challenges in longitudinal data analysis. In addition to inheriting all standard features of Gaussian processes, LonGP can model time-varying random effects and non-stationary signals, incorporate multiple kernel learning, and provide interpretable results for the effects of individual covariates and their interactions. We develop an accurate Bayesian inference and model selection method, and implement an efficient model search algorithm for our additive Gaussian process model. We demonstrate LonGP’s performance and accuracy by analysing various simulated and real longitudinal -omics datasets. Our work is accompanied by a versatile software implementation.AvailabilityLonGP software tool is available athttp://research.cs.aalto.fi/csb/software/longp/[email protected],[email protected]
Faculty Opinions recommendation of A review of the use of time-varying covariates in the Fine-Gray subdistribution hazard competing risk regression model.
