geckopy 2.0.2

Features:

• FSEOF implementation for ecModels in /geckomat/utilities/ (PR #119).

• Routines for quickly generating context-specific ecModels from a general ecModel, also in /geckomat/utilities/ (PR #120).

Fixes:

• Closes #116: Migrated CI to Travis-CI.com (PR #118).

• prot_abundance.txt can now be an empty file without the pipeline erroring (PR #124).

• Removes storage of metabolite notes in the ecModel, which were causing an error in COBRA (PR #125).

Documentation/Styling/Others:

• Added contributing guidelines, code of conduct and PR/issue templates (PR #115).

• Closes #123: Simplified GECKO pipeline log (PR #128).

Fixes:

• UBs are only changed for strictly required proteins (PR #103).

• Solves #101: All yeast models in the toolbox are built on the same yeast-GEM version 8.1.3 (PR #105).

• Solved bug in generate_protModels.m that constrained both the biomass and growth reactions (PR #107).

• Updated env. list after release of pandas 1.0 (PR #110).

• A folder for ecModels is added in case it doesn’t exist already (PR #111).

• Solves #112: Avoid using the variable name version for potential conflicts with Matlab (PR #113).

Style:

• Solves #102: All UBs in repo changed from +Inf to +1000 (PR #109).

Features:

• A main utility for constraining an ecModel with proteomics data stored in data files, and ensuring the model can grow at the provided conditions (PR #82).

• More constraining options in FVA utility (PR #95).

• An option for reducing display of fitGAM.m (PR #82).

• A chemostat simulation utility (PR #82).

Fixes:

• Fixes #73: Bug in getKcat.m (PR #75).

• Fixes #78: Updated BRENDA URL (PR #91).

• FVA utility: Fixed directionality bugs in MAX_min_Optimizer.m (PRs #69 & #95) & optimization bug in comparativeFVA.m (PR #95).

• geckopy: Updated csv reader in geckopy test (PR #83) and met id parsing (PR #93).

• Fixed subfolder bug in modifyKcats.m (PR #80).

Refactoring:

• Introduced getModelParameters.m for defining all input parameters, for streamlining the main ecModel generation pipeline (PR #76).

• Simplified constraining procedure in the FVA utility (PR #95).

Documentation:

• Improved documentation of several main functions (PR #76).

• Updated Matlab requirements (PR #96).

Features:

• Additional options for output tables from modifyKcats.m & topUsedEnzymes.m (PR #61).

• keggID is now an input for updateDatabases.m (PR #62).

• Backwards compatibility with any yeastGEM from 8.0.0 onwards (PR #66).

• New utilities:

  • getSubset_ecModel.m, for getting context-specific ecModels (PR #64).

  • getKcat.m, for retrieving kcats (PR #67).

Fixes:

• Fixed bug in aconitase kcat & misc. error handling (PR #62).

Refactoring:

• Speed improvements in topUsedEnzymes.m (PR #61).

• Reduced display of several functions (PR #62).

• Simplified changeMedia_batch.m and made more generic constrainEnzymes.m & flexibilizeProteins.m (PR #63).

Style:

• Changed EOL to LF (unix default) (PR #68).

Documentation:

• Documented input/output of topUsedEnzymes.m & truncateValues.m (PR #61).

• Added/updated documentation of changeMedia_batch.m, constrainEnzymes.m, flexibilizeProteins.m & getConstrainedModel.m (PR #63).

Features:

• Generalization of measureAbundance.m to receive any PaxDB file, a relative proteomics dataset, or even nothing at all (PR #58).

• New utility: Comparative FVA between a model and its enzyme-constrained version (PR #57).

Fixes:

• Consistent definition of what data is in uniprot.tab (PR #48).

• Proper use of measureAbundance.m from within constrainEnzymes.m (PR #56).

Refactoring:

• Switch all functions that add/change rxns/genes from COBRA to RAVEN (PR #48).

• Avoid any functions from Simulink (PR #48).

Fixes:

• Fixes #15: Binary results from the model (ecModel.mat, ecModel_batch.mat & enzData.mat) are no longer stored in repo (PR #52).

• Misc. fixes in the biomass composition + GAM calculations (PR #53).

Refactoring:

• Speed improvement in misc. functions (PR #49).

• Added sumProtein.m for easier use when creating new ecModels (PR #53).

Documentation:

• Documented better which scripts/data should be changed and which are optional when adapting geckomat to produce a new ecModel (PR #53).

Features:

• Name & version of the model are now read/stored from/as model fields (PR #42).

• Pipeline now works for any objective function (PR #47).

Fixes:

• Fixed bug from #39 that saved the .mat file with the wrong name (PR #42).

• Adapted pipeline to deal with multiple gene IDs for 1 protein / multiple protein IDs for 1 gene, for dealing with human-based GEMs (PR #43).

• changeMedia_batch.m modified to reflect the Y6 minimal media composition (PR #47).

Refactoring:

• Performance improvements to getConstrainedModel.m and sigmaFitter.m (PR #47).

• fitGAM.m is now only called from inside scaleBioMass.m (PR #47).

Features:

• Adapted the pipeline to work with yeast-GEM, including loading, processing and saving the model. Current model is constructed from yeast v8.1.3 (PR #39).

• When constructing ecModel_batch, lipid fraction is now scaled together with protein and carbohydrate fractions (PR #39).

Fixes:

• geckopy tests flexibilized to comply with yeast-GEM (PR #39).

Refactoring:

• Reorganized the repo, making a division between geckomat (Matlab part for generation + simulation of ecModels) and geckopy (Python part for simulations of ecYeastGEM) (PR #40).

• Parameters f (mass fraction of enzymes in model), Pbase, Cbase, Lbase (biomass composition) and GAM (growth-associated ATP maintenance) are now automatically computed (PR #39).

• Added RAVEN as a dependency for geckomat (PR #38).

• Changed most COBRA functions in pipeline to RAVEN functions (PR #39).

Features:

• Protein flexibilization: When proteomic measurements are provided, individual protein levels will now be iteratively flexibilized by the pipeline if the model results to be overconstrained, based on a provided growth rate. After this, flexibilized protein exchange pseudoreaction upper bounds will be set to the their flux values from a parsimonious FBA simulation (PR #34).

• Utilities: Included a folder with useful functions (PR #34).

Fixes:

• Fixes #14: CI is no longer failing, as model location, model naming and metabolite ID naming were corrected. test_adjust_pool_bounds was simplified to test with only 1 essential protein (PR #28).

Features:

• All genes from the original yeast model now included in the .xml file. Genes connected to enzyme constraints are now stored in model.enzGenes in the .mat structure.

• Docs badge in README.

Fixes:

• Fields grRules and rules fixed in a consistent way:

  • grRules for the backwards reactions are the same as for the forward ones.

  • For reactions catalyzed by just 1 enzyme (or complex), grRules of the original reactions are assigned to them.

  • For reactions catalyzed by more than 1 enzyme (or more than 1 complex), grRules of the original reactions are assigned to the arm reactions, and the corresponding sub-rules are assigned to the isozyme-controlled reactions.

  • For enzyme exchange reactions, grRules are assigned as thecorresponding gene ID.

  • The rules field is set equal to grRules for providing consistency with different toolboxes.

• Inter-OS compatibility: * Numbers in scientific notation are stored in the .xml files with format Xe-0N, not Xe-00N, or with format Xe-1N, not Xe-01N, regardless of the OS used for generating them. * Numbers in all files are shown with up to 6 significant figures.

Refactoring:

• Updated to new COBRA standards for addReaction usage.

NOTE: Not available in pypi (issue #14 unresolved)

Implemented automatic kcat flexibilization for over-constrained models:

• Based on a maximum growth rate specified by the user, the algorithm iteratively identifies the top growth-limiting kcat value and changes it for the highest one in BRENDA (same EC number)

• Once that the model is growing close to the set value, the average enzyme saturation factor is refitted

• For non-feasible/zero-growth models, sensitivity analysis is performed on a reaction and enzyme basis rather than on individual kcat values

• The outputs of this step are stored in topUsedEnzymes.txt and kcatModification.txt and can be used for further manual curation

All databases updated (BRENDA, swissprot, KEGG, PaxDB)

More generic gene/protein matching for compatibility with other models

Re-organization of all output files in a single folder

New badges + styling of website

NOTE: Not available in pypi (issue #14 unresolved)

Improved kcat matching to BRENDA with:

1. Specific activity

2. Phylogenetic distance, when data for organism of choice is not available

Switched to readthedocs for documentation: http://geckotoolbox.readthedocs.io

Added a Gitter room for discussion: https://gitter.im/SysBioChalmers/GECKO

Switched to a simplified GitFlow structure (master + devel + feature branches)

Python 3.4 environment dropped in CI (no longer supported by pandas)

NOTE: Not available in pypi (issue #14 unresolved)

Model and data are now also deployed.

Changes in license and readme.

First release on PyPI.

First release of GECKO in Github.