domains / families / science / trajectory-modeling

trajectory-modeling

separate

Models continuous temporal trajectories from BULK or time-resolved omics where the x-axis is measured experimental time: penalized GAMs (mgcv) for smooth trends and changepoint detection (segmented, ruptures) for abrupt regime shifts. Use when deciding between a smooth GAM and a changepoint model; choosing the GAM distribution (nb() plus a library-size offset for raw counts vs Gaussian on vst/log-CPM); setting the basis-dimension ceiling k below the number of timepoints and letting REML pick wiggliness; handling residual autocorrelation across timepoints with corAR1/bam(rho=); testing whether two conditions' trajectories diverge with an ordered-factor difference smooth; and choosing a changepoint search/cost/penalty (Pelt/Binseg, l2/rbf). Not for single-cell pseudotime (see single-cell/trajectory-inference).

When to use

• modeling continuous temporal trajectories
• penalized GAMs for smooth trends
• changepoint detection for abrupt regime shifts
• choosing between smooth GAM and changepoint models
• selecting GAM distribution for raw counts or transformed data
• setting basis-dimension ceiling k
• handling residual autocorrelation across timepoints
• testing trajectory divergence between conditions

When not to use

• single-cell pseudotime
• single-cell/trajectory-inference

Skill page

Source