PyClone-VI Released

[Andrew]

An alternative faster implementation of PyClone has been released. It is a slightly different model, and inference scheme. This has been done to better support WGS and exome datasets with many mutations.

If you are using a few hundred mutations there is probably no upside to the new version. But if you have more than say 500 it may be a lot faster. The downside is there is no plotting support, just a table output.

The paper is here https://www.biorxiv.org/content/10.1101/2020.08.31.276212v1 and the code is here https://github.com/Roth-Lab/pyclone-vi.

Would love to get some user feedback.

Best wishes,

Andy

[Zhuo Xm]

Dear Dr. Roth,

I tried the PyClone-VI on WGS samples with thousands of variants. It did perform much faster than PyClone (run time reduced from hours to mins).

Is all loci in one cluster going to have the same cellular_prevalence? (in PyClone, each loci have slightly different value)

mutation_id cellular_prevalencecellular_prevalencesample_id cluster_id cellular_prevalence cellular_prevalence_std cluster_assignment_prob 

chr11:49372870:T>C      A         0         0.6618        0.0113        0.4583 

chr11:49372870:T>C      B         0         0.7487        0.0109        0.4583 

chr12:100447011:C>A   A         0         0.6618        0.0113         0.4431 

chr12:100447011:C>A   B         0         0.7487        0.0109          0.4431 

chr12:108000697:A>G   A         0         0.6618       0.0113          0.4141 

chr12:108000697:A>G   B         0         0.7487       0.0109          0.4141 

best

Ming

[Andrew]

Yes that is expected. It is the cellular prevalence of the cluster the mutation is assigned to. In the original PyClone this value was the average of the MCMC trace which could be different then the cluster prevalence output by other commands.

[危志庭]

Can the  pyclone-vi support total copy number prior? or just can be used with major copy number prior?

[Andrew]

Just major. But if you set major cn to the total copy and minor cn to zero it is nearly the same as the total copy number prior.