Should be fairly easy :)

Transmart can have gene, probe or region level copy number data. cBioPortal supports gene and region (segmented, perhaps it could be used also for probe level).

For example, if my source file contains:

after running tm2cbio, the output will only contain the second one.

Given that cBio expects only one, we may have to aggregate them?

Currently tm2cbio labels every attribute in the data_clinical it generated as a sample attribute, even e.g. Race. Even though it is not currently an issues, as we have only 1 sample per patient, it would be better if a user can specify whether a clinical attribute is a patient or a sample attribute.

Desired solution:
allow user to specify the type in the mapping file

I'm playing around with the newmetadataformat branch and I generated clinical files for the cluc. In the most recent cbioportal release the clinical files were split into a Sample file and a Patient file. Hence, the datatype tag should now be SAMPLE_ATTRIBUTES
Furthermore, the line in the datafile which contains "SAMPLE" is no longer allowed, since this is already defined by the datatype tag.

The data_clinical.txt file created does not have an equal number of tabs in every line. It appears there are no tabs at all between the last non-empty field and the end of line.

This issue raises errors in the validator.

The tM data contains e.g. a column LOG2E, which I would like to use as data for my cBio import. My mapping file contains:

expression file path=data_mrna.tsv

one of VALUE, LOG2E or ZSCORE

expression data column=LOG2E

In the code, the data column is used directly as datatype
datatype: ${typeConfig.data_column}
However, cBio wants CONTINUOUS as its datatype. I think we're missing a mapping from data column to datatype, as now the datafile contains the correct data, but the meta file is wrong:

datatype: LOG2E
stable_id: null

I'm not 100% sure what's going on, but if my input clinical file contains:

PATIENT_ID RNAseq Illumina HiSeq2000\siNon-Targeting 1 RNAseq: Illumina HiSeq2000\siNon-targeting 1

This becomes

PATIENT_ID SAMPLE_ID siNon-Targeting_1 siNon-targeting_1

PATIENT_ID SAMPLE_ID RNAseq_Illumina_HiSeq2000/siNon-Targeting_1 RNAseq:_Illumina_HiSeq2000/siNon-targeting_1

STRING STRING STRING STRING

5 5 5 5

PATIENT_ID SAMPLE_ID siNon-Targeting_1 siNon-targeting_1

Even though I'll change the header anyway, apparently there is a bug here as the difference between:
RNAseq Illumina HiSeq2000\siNon-Targeting 1
and
RNAseq: Illumina HiSeq2000\siNon-targeting 1
is a ":" and a capital T and the difference we end up with is just the capital T. cBioPortal will not upload this, as it considers them the same.