ONP rendering breaks the OxoG filter (Part II -- field collapsing)

[LeeTL1220]

Please see issue #325 for more details. The first part of the OxoG filter fixes (issue #325 ) includes:

• TCGA MAF as input and output.
• Overwriting of existing annotations.

This issue will include the collapsing of number output fields as specified by user (with appropriate default). Particularly, this is useful for when we have DNPs and downstream tools expect a single number.

[LeeTL1220]

Please note that the collapsing will only happen when output is TCGA MAF. Please note that the current design could be leveraged for more versatile functionality in the future.

Collapsing will happen in the TCGA MAF output.

Now for a checklist:

• [x] Merge develop after pull request issue #327 is merged into develop.
• [x] Implement command line option to to collapse fields.
• [x] Create a default config file that is used in the case where no file is specified.
• [x] Build functionality that will honor the command line specification

Unit tests:

• [x] Create a mutation with fields to be collpased and fields that could be collapsed, but aren't. Make sure appropriate fields are combined.
• [x] Make sure config file can be read in and the proper collapsing list created.
• [x] Ensure that we treat float use case and integer use case properly. (could be rolled into test above)
• [x] Make sure that we can do the M2 workflow and see collapsed entries.

[LeeTL1220]

For the second Oncotator run in the pipeline.... As of this writing, the relevant task config in FH will be:

input_format: TCGAMAF output_format: TCGAMAF additional_options: --log_name oncotator_firehose.log --prepend --infer-onps -c /xchip/cga/reference/annotation/db/tx_exact_uniprot_matches.txt --collapse-number-annotations

No other task config parameters should be changed from standard (incl. no venv changes)

[LeeTL1220]

Here are commands that, as of this writing, I believe to work:

# Step 1 VCF to MAF (pre-oxoG filter)
oncotator -v -i VCF -o TCGAMAF --collapse-filter-cols --skip-no-alt test/testdata/m2_support/phasingExample.vcf ~/broad_oncotator_configs/phasingTest.vcf.maf.annotated hg19 --db-dir ~/dbdir_current -c ~/broad_oncotator_configs/tx_exact_uniprot_matches.AKT1_CRLF2_FGFR1.txt --log_name ~/broad_oncotator_configs/step1_oncotator.log --prepend

# Step 2 TCGA MAF to TCGA MAF with ONP and value collapse (post-oxoG filter)
oncotator -v -i TCGAMAF -o TCGAMAF ~/broad_oncotator_configs/phasingTest.vcf.maf.annotated ~/broad_oncotator_configs/test_reannotate.maf.annotated --infer-onps -c ~/broad_oncotator_configs/tx_exact_uniprot_matches.AKT1_CRLF2_FGFR1.txt hg19 --db-dir /home/lichtens/dbdir_current --prepend --collapse-number-annotations --log_name ~/broad_oncotator_configs/step2_oncotator.log

[LeeTL1220]

@kcibul Though this issue is closed, there is useful information for creating FH task configs/modules.