Cancer cell fractions instead of sequencing depth

[rahulk87]

Hi,

Thanks for this beautiful work. I was wondering if it make any sense (or possible) to run Treeomics with cancer cell fraction values (ccf) obtained from PyClone/ABSOLUTE?

Thanks, Rahul

[johannesreiter]

Hi Rahul,

Thanks for your question. It would not make much sense to use inferred CCFs with Treeomics for multiple reasons: i) a lot of signal to reconstruct phylogenies is lost due to the clustering of many mutations into individual subpopulations and hence Treeomics would not be able find the maximum likelihood solution as well as confidence values; ii) to find well-supported subclones Treeomics evaluates the variant reads and coverage supporting each variant and the number of variants supporting each mutation pattern - without these values a meaningful subclone detection would be impossible; iii) Sun et al (Nature Genetics, 2017) recently showed that inferred subclones often disappear when more regions of a cancer are sampled. Inferring phylogenies based on these non-robust subpopulations could be problematic which we try to avoid.

[rahulk87]

Thank you so much for your detailed answer.