How to quantify uncertainty for predictions made by timeOmics?

Hello,

I understand that at different parts of the timeOmics pipeline (e.g. checking optimal number of clusters using silhouette coefficient or checking distances within/between clusters), but I was wondering are there any additional ways to estimate uncertainty around the predictions that time Omics makes? For example, (if possible) using a resampling method like bootstrapping for the linear mixed model splines and getting confidence intervals or something else that might be more statistically valid. Any thoughts would be appreciated. Thank you

Hi @blueb,

Unfortunately we do not have estimates around the interpolation. timeOmics is based on our previous lmms package (see this paper A Linear Mixed Model Spline Framework for Analysing Time Course ‘Omics’ Data - PMC)

The lmms package is not maintained anymore, so if you want to have a look, it is better to use the cloud RStudio and choose the appropriate R version to load it). The main output of the lmms will give an idea of the type of model that was applied, but I don’t think it gives much more (to be checked).

Kim-Anh