Measuring tumor heterogeneity

Question

[ceruleanivy]

Does anyone know how it's possible to quantify heterogeneity between SNP mutations of the primary and metastatic site(s) of a tumor ? The goal is to possibly extrapolate about the origin of the metastatic legion and correlate a numerical index of heterogeneity with Overall Survival, Time for Progression etc. I've run across some examples but I don't know if they're applicable for what I'm asking.

http://www.ncbi.nlm.nih.gov/pubmed/23079694 http://www.ncbi.nlm.nih.gov/pubmed/26099045 (Figure 5a, Figure 4d)

[poisonAlien]

math score is the only way as far as I know, which can measure heterogeneity in terms of a numerical value.

Answer 1

[cbst]

I am not sure what kind of data you have, and I may have misunderstood your question, but here are anyway some tools that look at tumor heterogeneity, the clonality of the tumor and tumor purity, as well as evolutionary progression between several samples from the same patient. This list is probably incomplete.

Bayclone

THetA

SciClone

Clonet

CITUP

Battenberg

[ceruleanivy]

I have NGS data from IonTorrent with ref and var coverages, vaf. You might want to look at the above pubs, the Nature paper states something about generalized linear models and overdisperesed binomial families. The supplementary files provide actual scores for each mutation and patient.