Merge Two SEs

[KelseyChetnik]

In GitLab by @kelsey.chetnik on Jun 15, 2020, 13:11

Based on earlier concatMetabolon.R scripts and variations of it - don't know if this still applies?

[KelseyChetnik]

Create a merger function to combine two SEs. Must do the following:

• [x] accept one SE as the initial SE input, D1, and pass other as an argument, D2 (see mt_plots_stats_compare)
• [x] supply sample id cols for each SE as arguments; check that ALL sample ids match - crash otherwise
• [x] combine assay, rowData, and colData data frames
• [x] add a column to rowData differentiating the blocks (name it "block_[something]")
• [x] save metadata of D2 in output of merger function result list
• [x] make if-case for handling duplicated features

[KelseyChetnik]

Questions:

1. How handle case in which D1 and D2 have features in common?
2. How handle identical column names? Identical columns?
3. Should arguments for feature name columns in rowData be provided? Otherwise combining these might be completely disjointed. (Are we assuming that share columns in common? Keeping them separate?)

Assuming because intended for multiomics integration the first two won't typically be issues.

Answers:

1. make flag argument - crash_on_duplicate - set it to T by default. If crash_on_duplicate=F, add _1 and _2 to end of each duplicate. For colData - take from D1 (default) or D2 - we will assume if sample IDs must be identical, sample annotations should also be identical.
2. answered above
3. merge rowData where possible (ensure that the order is the same!)

[KelseyChetnik]

Implemented in 88d27d22b92265af7b760bb20cf98d6508d21ac7.