GECKO 3.2.2

• Fixes:
  • By default, the obj.params.uniprot.reviewed parameter in the adapterTemplate is set to false, to increase the number of matching proteins.
  • The model.ec.source entry after running setKcatForReactions has been corrected to 'setKcatForReactions'.
• Features:
  • If sensitivityTuning modifies a kcat value, it will now be indicated in model.ec.source as 'sensitivityTuning', and the previous kcat value and source will be kept in the model.ec.notes field as preTuneKcat=*value* | source:*original source*. If the kcat value had been subjected to senstivityTuning before, then the notes field will remain unchanged, so repeated runs of sensitivityTuning will not overwrite the notes.
  • runDLKcat support alternative paths and names of the DLKcat.tsv file (solves #392).
• Documentation:
  • Update references and add MATLAB badge in README.md.
  • Clarifies the meaning of obj.params.enzyme_comp in model adapter (solves #386).

GECKO 3.2.1

• Fixes:
  • getStandardKcat.m removed enzymes when it was run on a model already containing standard kcat values
  • ecFVA will swap min and max values the reaction directionality was swapped when mapping from ec- to conventional model
  • plotEcFVA will use its own cdfplot function, to avoid dependency on MATLAB Statistics and Machine Learning Toolbox. Blocked reactions are excluded from the graph. ⚠️ The plot in full_ecModel shows minor differences, also due to code changes introduced by GECKO 3.2.0
• Features:
  • calculateFfactor can handle protData with multiple conditions (in which case, averages over all conditions are taken)
  • setKcatForReactions allows multiple kcat values as input, allowing to set kcat of multiple reactions at the same time
  • reportEnzymeUsage will report both the summed usage of isoenzymes, as well as the individual usage of the separate reactions. topUsage and highCapUsage outputs columns are harmonized
  • constrainFluxData if a provided constraint is either -1000 or 1000, lower and upper bound will be constraint to either uptake or excretion only, irrespective of what option is given as the looseStrictFlux parameter
  • constrainEnzConcs can also remove enzyme concentration constraints
• Refactor:
  • fuzzyKcatMatching to avoid warnings "Colon operands must be real scalars" in recent MATLAB versions
• Documentation:
  • various function documentations, warnings and error messages have been improved

GECKO 3.2.0

• Fixes:
  • ⚠️ specify RAVEN 2.9.2 as dependency, and provide more detailed instructions on how to upgrade
  • loadDatabases clarifies what to do if download from UniProt fails
  • calculateFfactor sometimes failed to properly read paxDB.tsv files
  • runDLKcat should parse params.path as absolute paths to docker #378
  • applyKcatConstraints can handle when all kcat are zero
  • getStandardKcat should also assign standard kcat values for reactions that do have a non-empty grRules field, but are not in model.ec due to the genes not matching any of the enzymes in the uniprot.tsv file
  • calculateFfactor can take protData input
  • enzymeUsage correct mention of output units #376
  • updateProtPool is made obsolete #375
• Features:
  • ⚠️ keep protein usage reaction draw from protein pool when proteomics is integrated #375
  • new removeStandardKcat function removes all traces of getStandardKcat having modified a model
• Chores:
  • update dependencies in GitHub Actions

GECKO 3.1.3

• Fixes:
  • resolve a bug when saving the ecModel in xml format #361
• Documentation:
  • update DOI in code and Readme #362
• Features:
  • clearer error message for use of Docker in Matlab #363

Full Changelog: v3.1.2...v3.1.3

GECKO 3.1.2

• Fixes:
  • makeEcModel prevent duplicated protein pseudometabolites
  • applyKcatConstraints occassionally fills S-matrix with NaN in light ecModels (solves #344)
  • calculateFfactor handle data from PAXdb if the taxonomic ID is not 4 digits long (solves #345)
  • runDLKcat correctly handles param.path if the folder name has spaces (solves #351)
  • loadDatabases throws error if duplicate protein IDs are found
  • getSubsetEcModel requires both bigEcModel and smallGEM to have derived from the same starting GEM, and will therefore check whether all reactions from smallGEM are also present in bigEcModel. (solves #353)
• Documentation:
  • Installation instructions are moved to the Wiki
  • Mention correct human-GEM version in HumanGEMAdapter.m
  • Replace Gitter with GitHub Discussions for asking support
• Refactor:
  • Have ecFSEOF follow the original implementation of FSEOF (PR #356)
• Features:
  • Navigate into new project folder after running startGECKOproject

GECKO 3.1.1

• Features:
  • addNewRxnsToEC to add new enzyme-catalyzed metabolic reactions to an ecModel. (PR #337)
  • readDLKcatOutput reports which metabolites or reactions are not matching. (solves #334)
  • mapRxnsToConv throws error if empty flux vector is used as input. (solves #332)
• Fixes:
  • getComplexData does not write empty complex data. (solves #338)
• Refactor:
  • Rename Prot to Enz to give the new function names flexibilizeEnzConcs, fillEnzConcs and constraintEnzConcs.
  • ecFSEOF as one combined function to run FSEOF simulations on ecModels.
• Documentation:
  • Minor edits in protocol.m files.

GECKO 3.1.0

Main improvements in this PR:

• Features:

  • Renamed various parameters in model adapter.
  • getComplexData uses taxonomic ID instead of species name.
  • GECKOInstaller checks if correct RAVEN version is installed (release 2.8.3+).
  • Tutorials are moved to tutorials, and userData is removed as default location for model files. Users are encouraged to make their model folders outside the GECKO directory, facilitated by startGECKOproject().
  • full_ecModel tutorial:
    • more detailed discussions around kcat curation, protein usage etc.
    • plotCrabtree makes plot that demonstrates Crabtree effect in ecModels
    • plotlightVSfull makes plot comparing flux distribution in light and full ecModels
  • light_ecModel:
    • minimum code to make a light ecModel based on human-GEM 1.15.0
    • make contextualized ecModel based on a cell-line specific tINIT model with getSubsetEcModel
  • ecFSEOF is compatible with GECKO3.
  • makeEcModel can use pseudoRxns.tsv to filter out pseudoreactions.
  • various functions show progress bar that indicates estimated remaining time.
  • reportEnzymeUsage can make report of top-10 used enzymes.
  • loadConventionalGEM can load yaml model files.
  • enzymeUsage reports as positive values.
  • flexibilizeProtConcs also keeps track of the ratio of protein concentration change.
  • fillProtConcs allows for selection of column from protData, if it contains multiple datasets.
  • setKcatForReactions can directly modify a kcat for all associated reactions, useful for manual curation on existing ecModels.
  • getReactionsFromEnzyme gives which reactions are catalyzed by a particular enzyme, from a provided Uniprot ID.
  • copyECtoGEM to copy ecModel.ec.eccodes to ecModel.eccodes.

• Documentation:

  • All function documentation is available from /doc/ and online (becomes available upon PR commit).
  • sensitivityTuning allows for ignoring selected reactions.

• Fix / refactor:

  • Numerous fixes and refactoring of code in many functions.
  • Note that various functions have had their input parameters changed.

GECKO 3.0.2

• Documentation:
  • Clarifications and fixes in protocol.m (solves #279, #287).
  • Protein concentrations (in model.ec.concs and proteomics data) are in mg/gDCW.
• Fix:
  • Correct calculation of new protein pool if constrained with proteomics, by considering the total amount of measured protein.
  • getStandardKcat can use data/pseudoRxns.tsv as input of reactions to ignore, it overwrites previous standard Kcat definitions, and specifies model.ec.source{i} = 'standard' if a standard kcat is assigned to a zero-kcat reaction (solves #280, #286)
  • Various bugfixes in fuzzyKcatMatching (solves #277), readDLKcatOutput (solves #278), getECfromDatabase (solves #281), sensitivityTuning (solves #285), & applyCustomKcats (solves #291).

GECKO 3.0.1

• Fix:
  • getECfromDatabase incorrect definition of ecRxns and action input parameters.
• Features:
  • flexibilizeProtConcs prevent infinite runs in no (more) proteins are limiting and check if protein pool is constraining growth.
• Documentation:
  • Updated Issues and PR templates.
  • Updated contributor guidelines.
  • Updated README.md.
  • Updated protocol.m.
  • GitHub Pages.

GECKO 3.0.0

After significant refactoring of the codebase, GECKO 3 is more user-friendly, flexibile and versatile than before.

Most notable changes:

• ecModels have an .ec structure containing all enzyme and kcat information.
• Enzymes are incorporated in the S-matrix as MW/kcat (previously this was 1/kcat, where the MW was instead considered in the protein exchange reactions).
• ecModels can be stored in YAML file format that retains all model content.
• The model-provided list of EC numbers can be used.
• Various kcat sources can easily be combined.
• DLKcat is distributed as part of GECKO, as alternative kcat source.
• Enzyme complex data can be gathered and considered.
• All model-specific files and scripts are kept in the userData subfolder.
• Model-specific modelAdapter files contain parameters that are used in model reconstruction and analysis.
• Because of the above, there is virtually no backwards compatibility with GECKO versions 1 and 2, regarding both models and code.