biodavidjm / artms Goto Github PK

• README
• Vignettes
• Source code comments

2do:

• artms_enrichForComplexes should break if anything happens to artms_data_corum_mito_database
• All the functions used across the repository, as for example, artms_changeColumnName

Currently, there are two independent functions to generate the MIST format based on spectral counts or intensity. Unified both

It might require too much time, so not sure if worthy. However, in the long term...

Why do you need so many global variables?

Should it be removed? if no solution is found for the large number of files that are necessary, removed.

A reasonable solution could be: download the data from kroganlab website.

• Remove aggregate_fun option

• Break each of the large functions in multiple files
• Add comments
• Fix warnings
• Remove MSstats_main

Right now is reading from the system. But this is wrong.
It should use artms_annotation function

When selecting the output name by the artms_proteinToSiteConversion tool, prevent that the output from having the same name as the original file (i.e., overwrite)

• The data option should include all the data related options (silac, filters, fractions)
• Rename files$data to files$evidence
• Add qc section
• Eliminate unused options

When the server is down, it throws the following error:

 Error in function (type, msg, asError = TRUE)  : Empty reply from server 

which breaks artms_analysisQuantifications

Currently is pointing to the wrong local file. Make it available in the package

When running artms_quantifications, we need a second options in qc to run the extended functions

• Fix bug with respect to the output name of the function artms_generatePhSiteExtended
• Check if ptmsites is the right option based on the PTM_ph or PTM_STKY notation

WARNING: Use TRUE/FALSE instead of T/F
Found in R/ directory functions:

• .artms_checkIfFile()
• .artms_EnrichmentPlotHeatmaps()
• .artms_foldComplexEnrichment()
• .artms_getMSstatsFormat()
• .artms_getPCAplots()
• .artms_imputeMissingValues()
• .artms_plotHeat()
• .artms_read_evidence_file()
• .artms_removeMaxQProteinGroups()
• .artms_runMSstats()
• .artms_sampleCorrelationHeatmap()
• .artms_samplePeptideBarplot()
• .extras_annotate()
• .select_ref()
• artms_analysisQuantifications()
• artms_annotationUniprot()
• artms_avg_intensity_RT()
• artms_dataPlots()
• artms_enrichLog2fc()
• artms_enrichProfiler()
• artms_evidenceQC()
• artms_evidenceToMISTformat()
• artms_evidenceToSaintExpressFormat()
• artms_filterEvidenceContaminants()
• artms_generatePhSiteExtended()
• artms_isEvidenceNewVersion()
• artms_mergeEvidenceKeysByFiles()
• artms_msstats_summary()
• artms_plotCorumEnrichment()
• artms_plotHeatmapQuant()
• artms_proteinToSiteConversion()
• artms_quantification()
• artms_replicatePlots()
• artms_resultsWide()
• artms_SILACtoLong()
• artms_spectralCounts()
• artms_volcanoPlot()
• artms_writeContrast()

Found in files:

• inst/extdata/createData.R
• man/artms_annotationUniprot.Rd
• man/artms_mergeMaxQDataWithKeys.Rd
• man/artms_volcanoPlot.Rd

It is a key information that should also be provided

Something is wrong when there is only one comparison for the heatmap> the heatmap gets way off.

Enrichment.plotHeatmaps

We need more extensive quality control check from the evidence and the summary file

Some functions might need to increase the number of controls to prevent errors

Break down MaxQ_utilities in individual files.
When naming the individual scripts, try to organize it based on major functions.

It is annoying to deal with the RawFile column name everywhere. Write a function to check and return the dataset (evidence or keys) with the RawFile column name

The ggalt library cannot be installed on linux. Therefore, it has to go away from the QC extended functions. This is the error:

*** Install PROJ.4 and if necessary set PKG_CPPFLAGS/PKG_LIBS accordingly.
ERROR: configuration failed for package ‘proj4’
* removing ‘/home/travis/R/Library/proj4’
Installation failed: Command failed (1)
missing: ggalt, proj4
The command "Rscript -e 'deps <- devtools::dev_package_deps(dependencies = NA);devtools::install_deps(dependencies = TRUE);if (!all(deps$package %in% installed.packages())) { message("missing: ", paste(setdiff(deps$package, installed.packages()), collapse=", ")); q(status = 1, save = "no")}'" failed and exited with 1 during .

Removing ggalt only affects QC-IntCorrelation.pdf
#40

I'll take care of this @alexproteomics

Currently it does a correlation plots taking all the features. Add an option to select the ptm type

Post-processing of msstats results, including:

• Set org.db as the resource for annotations and mappings
• Remove unnecessary steps
• Imputation
• Plots
• Basic Enrichment analysis
• Summary excel file

Issue to keep track of all the sample data objects created for the package as a way to provide compelling use cases for the package’s functions.

The first one is the PH dataset provided by Danielle

The input file requires by SAINTq is very different from SAINT express. Create a function to generate the input file from the evidence + keys (an additional column will be required: SAINT)

Yes, move annotation packages to Suggests. Otherwise, the user is required to install all of those packages that they may not need.

Your artms_mapUniprot2entrezGeneName function should be using a data.frame map that has two columns: common name and the annotation abbreviation (e.g., "Anopheles", "Ag"). Then you can use this along with if (!requireNamespace()) then stop("Install to map IDs") to get the user to have the package already installed before using your function. It would also reduce the repetitive code.

You could have used a data.frame map for the 'mapUniprot2entrezGeneName'
function. Is the function necessary?

• Plot based on fold enrichment instead of pvalue
• Add false discovery rate

pathogen might be needed to be removed. Otherwise is too complicated. Not sure how many people are mixing human and pathogens.
If used, where is the info about pathogens coming from? The user should provide the annotation file, but of course, it would have to be specified how it should be formatted (or how to get it)

Jittered plot would have to be removed as well

why is only human or mouse the only two species supported when using org.db there are way more available. Revise this.

It is a more intuitive name

It could confused the users to document functions that are meant to be used only internally

Generate input files for PHOTON from -imputedL2fcExtended.txt

Many of the functions should be use only internally:

Prefix internal functions with a '.'. Do not @export and in general skip royxgen docs for these functions, with the exception of @importFrom lines.

Give the user the option to provide a file name output, otherwise, use a build-in default one

• Add functions for QC analysis of the MaxQuant evidence file
• Add options to the configuration file to enable QC

• Add comments
• Remove deprecated functions
• Rename

I meant that your if statement should be more like
any(is.missing("datafilearg"), is.missing("textfilearg")) instead of testing with all(is.null(...), is.null(...)).
The logic in these tests is slightly different and perhaps you don't want any of the file
arguments to be missing rather than the instance where all of them are missing.

It includes:

• Help pages
• Vignettes
• Unit tests
• data/
• NAMESPACE
• CI: code coverage: codecov
• CI: travis
• CI: AppVeyor (dimissed, it does not support bioconductor packages)
• Do not include Packrat: isolate the packges installed for this project. Possible problems with bioconductor

Should be part of the same function. File plotHeatmap

General tasks related to the old functions that need to be transformed and get package format

• Clean up the functions
• Add unit tests
• Remove unnecessary functions (recommended at the end)

All the functions' arguments must incorporate the type in the documentation.

Make whatever is necessary to make the 80% of the functions runnable

Do it for this functions:

• R/evidenceQC.R:artms_evidenceQC
• R/avg_intensity_RT.R:artms_avg_intensity_RT
• R/MSstats_functions.R:artms_changeColumnName
• R/MSstats_functions.R:artms_filterMaxqData to artms_filterEvidenceContaminants
• Merge R/MSstats_functions.R:artms_mergeMaxQDataWithKeys and artms_mergeEvidenceKeysByFiles
• R/MSstats_functions.R:artms_resultsWide
• R/MSstats_functions.R:artms_spectralCounts
• R/MSstats_functions.R:artms_writeContrast -> internal
• R/annotations.R:artms_annotationUniprot
• R/annotations.R:artms_mapUniprot2entrezGeneName
• R/enrichments.R:artms_enrichForComplexes -> make internal
• R/enrichments.R:artms_enrichLog2fc
• R/enrichments.R:artms_plotCorumEnrichment -> make internal
• R/enrichments.R:artms_enrichProfiler
• R/plots.R:artms_plotHeatmapQuant
• R/volcanoPlot.R:artms_volcanoPlot
• R/replicatePlots.R:artms_replicatePlots
• R/checkIfFile.R:artms_isEvidenceNewVersion
• R/analysisQuantifications.R:artms_annotateSpecie

All these functions cannot have runnable examples due to either size restrictions or run for too long:

Long Running time

• artms_quantification: takes too long
• analysisQuantifications.R: takes too long

Require extra (large) files

• artms_SILACtoLong: requires a SILAC evidence file
• artms_evidenceToSaintExpressFormat: requires an APMS dataset
• artms_evidenceToMISTformat: requires an APMS dataset
• artms_msstats_summary: requires a summary file
• artms_dataPlots: requires uploading an extra file
• artms_generatePhSiteExtended: requires large extra files

Taking as input -imputedL2fcExtended.txt

As an R user, I have always liked to better know the functions available in every package.

An elegant solution would be to just providing a prefix with the name of the package

artms_NAME_OF_THE_FUNCTION

RMSQv3 works accepting flags/options. Remove everything related to library(getopt)

Problem with select() returning many:many mapping between keys and columns. Fix it. Find out how to get the primary gene symbol

just print out the config.yaml file to be filled up by the user

Currently the log2fc value that can be obtained after imputation might be too high if the intensity in the condition where the protein was consistently found is too high. Normalized the imputed values using the maximum of the log2fc calculated by MSstats.

Suggestion: provide two new arguments with the highest and lowest log2fc values to adjust the imputation method