Clonal Evolution Plot (figure 3d)

Question

[david.melnekoff]

Hello,

I'm having difficulty finding any information about how figure 3d, the schematic showing Clonal Evolution, was generated. Any information would be greatly appreciated!

Thanks -David

[GenoMax]

Where can one find that figure 3d?

[david.melnekoff]

in the sciclone PLOS paper

[GenoMax]

You had better provide a link in your original post since people (including me) are not going to go searching for it.

Answer 1

[Chris Miller]

You're referring to the plot in Figure 3d here: http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003665

That was hand-generated in Illustrator, but you have fortuitous timing. I'm just wrapping up an R package to generate those plots: https://github.com/chrisamiller/fishplot

Give it a try, and I'd welcome feedback on that early release.

[GenoMax]

Great timing on @david's part indeed.
And you saved him the effort of pasting a link to the paper as well.

[david.melnekoff]

Great! I'll give it a try and let you know.

[david.melnekoff]

Hi Chris I have one more question. Does Sciclone output tumor fraction of clones in any indirect or direct way? Or is it as simple as : # of variants in cluster one, over total number of clustered variants would give clonal fraction of clone#1?

Thanks again, David

[Chris Miller]

I'd recommend feeding the sciclone results into clonevol, which will infer phylogeny and give clonal fractions. https://github.com/hdng/clonevol

[cbst]

I am using Clonevol, and it is giving me two or more models for the same patient, for example one with model probability of 0.49 and one with model probability 0.47. I only have two samples per patient. How to you choose the model when it gives several possible models. Is it enough to choose the one with the highest probability?

[Chris Miller]

If there are two models with high probability, then it means there's not enough information in the data to distinguish between the two. There's likely nothing you can do without getting more data (in the form of another sample that's spatially or temporally distinct)