search

zanmer / Mesothelioma_evolution_deciphering_drugable_somatic_alterations

Analysis scripts and relative data used in the paper "Clonal Architecture in Mesothelioma is prognostic and shapes the Tumour Microenvironment" by Zhang et al.
3 stars 2 forks source link

Something about citup #2

Open mugpeng opened 3 years ago

mugpeng commented 3 years ago

Hello,

I notice that you did phylogenetic trees in CITUP with the output from PyClone. However, In my work, run_analysis_pipeline commands can't give an output of mutations which didn't appear in all samples. Which means, I can only make phylogenetic trees for the public mutations. image

In your paper, I find you do make the cluster of private mutations, and also do citup. How can you make that with no private mutations result from pyclone?

Below is my hypothesis: You just classify each groups that are not public mutations as a specific cluster, and for the input file of run_citup_qip, you just use the formal mutations frequency data, and in the sample with no mutations, just put 0. image image

Thanks.

zanmer commented 3 years ago

For multiple samples pyclone analysis, I have used bam-readcount to retrieve ref and alt depth of somatic mutations from samples that not detected these mutations. This process was usually called "force calling". After force calling, private or branched mutations show 0 depth for alt allele in some samples. This is necessary for multiple samples to run pyclone. Otherwise, only mutations present in all samples will be clustered.

mugpeng commented 2 years ago

Thanks zanmer.

I used the mentioned force calling. However, I got pretty low ccf from pyclone which the max is only around 0.6. Where you got the ccf in that picture which the ccf of overlap mutations are 1 ?

Besides, I got much more cluster in my results taking very long time for running run_citup_qip (15 parallels).