If only one out of five assays has replicate samples, is it possible to avoid specifying a sampleMap row for every column in every assay? Also, I notice assay columns not in colData are removed. How should samples which are technical controls, such as the pooled sample in every set of 8 iTRAQ labelled proteomics be kept? Is a dummy row in the patient clinical data table necessary?

Hi Dario,

Is it possible to avoid specifying a sampleMap row for every column in every assay?

No, we track all the columns in the assays in the sampleMap.

We recommend taking care of the replicates with mergeReplicates or by separating them into their own assay with splitAssay.

Also, I notice assay columns not in colData are removed.

We don't support orphaned samples in the data. Every sample should have a corresponding biological unit in the colData.

How should samples which are technical controls, such as the pooled sample in every set of 8 iTRAQ labelled proteomics be kept?

This depends on what operations you want to carry out. Without knowing much about this data, I would say that having an assay for the technical controls usually works.

Is a dummy row in the patient clinical data table necessary?

Yes, under our API, every assay column should have a corresponding row in the colData. This is a 1-to-many match and we do not have support for a many (colData rows) to one (assay column) match, if I am understanding your question correctly.

Best,

Marcel