While working on the BitPhylogeny paper, we stumbled on the problem of how to compare trees of clonal evolution.
Clonal evolution trees combine a clustering of molecular markers with tree inference. There are methods to compare clusterings and methods to compare trees, but how do you compare both at the same time?
Here is how we did it:
Do you remember the first post in this series, where we stated the intra-tumor phylogeny problem? No worries, if not – here it is again: Given a sample of the genomes of a heterogeneous tumor, identify the genetic clones and infer their evolutionary relationships.
Finally it’s time to announce our own approach to this problem, which has just come out in Genome Biology:
Yuan*, Sakoparnig*, Markowetz^, Beerenwinkel^. (2015). BitPhylogeny: a probabilistic framework for reconstructing intra-tumor phylogenies. Genome Biology 2015, 16:36
BitPhylogeny stands for `Bayesian inference for intra-tumor phylogenies’ – I am sure we mention this in the paper somewhere but am hard pressed to put my finger on where exactly this is. Well, now you know. Links to the code are for example here on my webpage.
How does BitPhylogeny work?