search

kellrott / SMC-Het-Challenge-Eval

0 stars 1 forks source link

SVclone and CCube #7

Open galder-max opened 5 years ago

galder-max commented 5 years ago

Ke will update the dockers for these and come back to us.

galder-max commented 5 years ago

Update from Ke:

_The updated Ccube/SVclone docker can be find here.

Updated: https://github.com/keyuan/docker-ccube/tree/smc-het-dev Original: https://github.com/keyuan/docker-ccube/tree/smc-het-old

There are two versions: “ccube.xml” (using Ccube purity) and “ccube_bb.xml” (Using Battenberg purity)

The major differences are the following

  1. Ccube/SVclone now handles variants in clonal and subclonal CNs jointly during clustering.
  2. There is a filter for false positive (CCF = 0) variants. _
kellrott commented 5 years ago

Galaxy invoked the command 

run_analysis_ccube_bb.R /export/galaxy-central/database/files/000/dataset_3.dat /export/galaxy-central/database/files/000/dataset_1.dat /export/galaxy-central/database/files/000/dataset_2.dat

It used the working directory

/export/galaxy-central/database/job_working_directory/000/39/working

Which resulted in the error:

Fatal error: Exit code 1 ()

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

Loading required package: methods
Loading required package: foreach
Loading required package: iterators
Loading required package: parallel
sh: 1: -v: not found
Error in system(shellCommandMutectSmcHet, intern = TRUE) : 
  error in running command
Execution halted
keyuan commented 5 years ago

It seems the script wasn't calling the vcf parser. I've just pushed a fix https://github.com/keyuan/docker-ccube/commit/692ab447115332ed5c400d3c1b5435f4a5e75a0d

kamichiotti commented 4 years ago

I'm having a problem with ccube running all of the tumors. I have 57 successful runs and then 17 that just won't run properly. I did have to make some modifications to the Dockerfile in order to get the first 57 to run, because the image didn't include the run_analysis_ccube_bb.R script nor a number of the functions available in the current version of keyuan/ccube on GitHub, such as ParseSnvCnaBattenberg. So, I made a new Dockerfile building off of keyuan/docker-ccube and installing keyuan/ccube from GitHub and a copying in run_analysis_ccube_bb.R. Strangely, I just cannot get these last 17 to run. I end up with the following error:

Error in { :
      task 1 failed - "no method for coercing this S4 class to a vector"
Calls: RunCcubePipeline -> %dopar% -> <Anonymous>
Execution halted
ESC[1;30mINFOESC[0m [job ccube.cwl] Max memory used: 0MiB
ESC[1;30mWARNINGESC[0m ESC[33m[job ccube.cwl] completed permanentFailESC[0m
ESC[1;30mWARNINGESC[0m ESC[33mFinal process status is permanentFailESC[0m

I'll email you an input tumor file as well as some of the corresponding outputs to help in debugging.

Thanks!