Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 22 additions & 0 deletions data/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
# Data

| File | Contents |
|---|---|
| `abs_proteomics.txt` | Absolute protein abundances [mmol/gDW]. 2929 rows, one header; three biological replicates per condition and four for hGR. Missing measurements are `NA`. |
| `fermentationData.txt` | Per condition: total protein content [g/gDW], dilution rate [1/h] and measured exchange rates [mmol/gDW/h]. An undetected byproduct is `NA`. |
| `ribosome.txt` | Ribosomal subunits with gene names, mass and sequence, from UniProt. |
| `selectedAnnotation.txt` | Assignment of proteins to the systems used when summarising enzyme usage. |
| `uniprot.tsv` | Yeast proteome with molecular masses, used to convert abundances from mmol/gDW to mg/gDW. |

## abs_proteomics.txt

`P61830` appears on two rows, once per gene of the duplicated histone
(`YBR010W`, `YNL031C`), with identical abundances. Readers collapse it to one
entry and fail if the duplicated rows ever disagree.

## uniprot.tsv

From [SysBioChalmers/GECKO](https://github.com/SysBioChalmers/GECKO),
`tutorials/full_ecModel/data/uniprot.tsv`, branch `develop4`, commit
`cbc4ca3311df8ca44c28792f40008b9cbef692bd`. UniProt proteome UP000002311,
reviewed entries. Covers every protein in `abs_proteomics.txt`.
Loading
Loading