Refactor workflow cbioportal_export

[eudesbarbosa]

Suggested changes

• [ ] All class names should use CapWords convention.

• [ ] Method cbioportalExportStepPart.iterate_over_biomedsheets should be broken up to follow single-responsibility principle:

  • iterate_over_biomedsheets(): Responsible for yielding biomedsheets contents
  • cbioportalCnaFilesStepPart._get_input_file_gistic(): Responsible for input files for action 'gistic'
  • cbioportalCnaFilesStepPart._get_input_file_log2(): Responsible for input files for action 'log2'
  • cbioportalCnaFilesStepPart._get_input_file_segments(): Responsible for input files for action 'segments'
  • cbioportalZscoresStepPart._get_input_file_zscores_input(): Responsible for input files for action 'zscores_input'

Alternatively, iterate_over_biomedsheets could be run once in the workflow constructor, the information could be stored in a dictionary for instance. Examples from VariantExportWorkflow._build_ngs_library_to_kit(): code

• [ ] In cbioportalExportStepPart.iterate_over_biomedsheets, both {mapper} and {tools} should be extracted from config instead of being hardcoded.
• [ ] Targeted CNV vs. WGS CNV relation should be made explicit. Currently, it is defined by tool listed in config. The step will mangle or omit CNV results if there is WES & WGS data in the project.

      if self.config["path_copy_number_step"]:
          if self.config["cnv_tool"] in ["cnvetti_on_target_postprocess", "copywriter"]:
              self.register_sub_workflow(
                  "somatic_targeted_seq_cnv_calling",
                  workdir=self.config["path_copy_number_step"],
                  sub_workflow_name="copy_number_step",
              )
          else:
              self.register_sub_workflow(
                  "somatic_wgs_cnv_calling",
                  workdir=self.config["path_copy_number_step"],
                  sub_workflow_name="copy_number_step",
              )

• [ ] Snakemake file should be modified so all rules get their output and input from python part.
• [ ] All rules in Snakemake should have a log file associated.
• [ ] Merge rules with similar outputs (e.g., cbioportal_export_CNA_log2 and cbioportal_export_CNA_calls.
• [ ] Consider merging rules that are simple shell calls (simple post/preprocessing?) - e.g.: cbioportal_export_concatenate_maf.

[eudesbarbosa]

@ericblanc20 and @messersc, for future references - this issue.

[eudesbarbosa]

Extracted from #156 :

Based on commit c46917387b3552c08113b6fab9e88e88c18c6ebf

• [ ] Update cbioportal_export default config.
• [ ] Fix wrapper snappy_wrappers/wrappers/cbioportal/case_lists/wrapper.py..
• [ ] Fix wrapper snappy_wrappers/wrappers/cbioportal/meta_files/wrapper.py.
• [ ] Fix wrapper snappy_wrappers/wrappers/vcf2maf/vcf2maf/wrapper.py.