Skip to content

Advanced Workflows

As part of my thesis at ETH Zürich I, gordonkoehn, designed a series of advanced Snakemake workflows as reproducible experiments.

We investigated particular aspects of SCITE’s MCMC inference:

  • warm-up of the MCMC chain
  • multimodal posterior distributions
  • convergence diagnsotics in the mutation tree space

The workflows are available in:

  • mark01 : assessing the HUNTRESS trees with distance metrics under the SCITE generative model

  • mark02 : investigate convergence of SCITE MCMC chains, given different initial points with tree distances.

  • mark03 : investigate convergence of SCITE MCMC chains, given different initial points with tree distances.

  • mark04 : assessing the HUNTRESS trees with distance metrics under the SCITE generative model.

  • mark05 : assessing the MCMC tree dispersion per tree-tree metric as a baseline to understand the tree-tree metric.

The outcome and analysis of the results are available in the thesis.