vitek-lab/msstatsconvert

Converter functions for the MSstats package

MSstats development version

MSstats is an R-based/Bioconductor package for statistical relative quantification of peptides and proteins in mass spectrometry-based proteomic experiments. It is applicable to multiple types of sample preparation, including label-free workflows, workflows that use stable isotope labeled reference proteins and peptides, and workflows that use fractionation. It is applicable to targeted Selected Reactin Monitoring(SRM), Data-Dependent Acquisiton(DDA or shotgun), and Data-Independent Acquisition(DIA or SWATH-MS). This package implements converter functions that are used by MSstats to import data from various signal processing tools.

Installation

This development version is only available on Github:

devtools::install_github("Vitek-Lab/MSstatsConvert", build_vignettes = TRUE)

Documentation and examples

An extensive documentation can be found on the official website of the package.

Contributing

We welcome contributions from the community. For details on how to contribute to the development of MSsstas, please refer to the CONTRIBUTING file.

License

Artistic-2.0

msstatsconvert's Issues

Handle Time column in annotation

add it to Condition column with a double underscore separator

@huang704

MaxQuant, PD TMT output: handle ratios of intensities

TMT converter: remove the parameter rmPSM_withMissing_withinRun = FALSE

Remove the parameter rmPSM_withMissing_withinRun = FALSE and all related codes from TMT converters

Shared peptides support

On-disk backend

Make balanced design before returning data

particularly in TMT converters

Output formats changed: PD, PDTMT, Spectronaut

PD:

• abundance column (name)
• protein ID column (name)

PDTMT:
(above? +)

• useProteinsNumCol (name)

Spectronaut:

• additional columns for filtering
• remove unnecessary columns

Remove unncessary warning messages

Conflicting annotations

Add small example input files for tests

remember about size restrictions for the package

Function to help work with MSnbase classes?

Check column names and availability in each converter

Handle multiple columns of the same name in annotation

for example as in DDA-Meierhofer2016 benchmark dataset for MaxQuant

Try to separate / remove multiple protein IDs in the same row

this happens in some converters, for example PD - there should be an option to remove rows with multiple protein IDs

Initial data report function

Fix names of columns in SpectronauttoMSstatsFormat

ProteinName in features for fractions/balanced design

yes or no?

Add an option to remove protein groups in all converters

The separator can be different for different tools.
Shared peptides here?

Check this when validating the results and fix

Account for annotation source

If annotation is taken from the input, remove Condition and BioReplicate columns or make a flag for this

convert checks to assertions

LF converter : fewMeasurements

LF converter : fewMeasurement option -> removeFewMeasurements or removefeaMeasurement and TRUE/FALSE.

MSstatsBalancedDesign function

fractions + fill incomplete rows

Add smaller functions for common parts in TMT/label-free converters

Make sure channels match annotation with the current version of .standardizeColnames

Check the value of fix_missing parameter in MSstatsBalancedDesign for each converter

before release

Move the documentation for exports from MSstats/MSstatsTMT to here?

Remove unnecessary df names when subsetting dt

Remove underscores from PeptideSequence and maybe Condition

Early in the MSstatsPreprocess function

@huang704

Examples in documentation for each exporter function

Function for conversion from wide format to long format

it will be helpful

• it should solve some of the possible problems with software versions and column names for data in wide format

Compare methods of aggregation

in summarize multiple measurements
and aggregate isotopic peaks
dcast vs current

Way to handle changing software versions (=input file formats)?

Check discrepancy of No. features and rows for LF

Check gSheet for LF

Logging setup

TODO: this should be happening on load, because we don't know yet if the user want to append?
OR if we don't save at this moment, the file won't be created, yet - file name can go to options, too?

OpenMS converter for TMT?

Make sure that duplicated biorep IDs are found in annotation

Vignette

• describe, how to use MSstatsPreprocess with data not coming from any of the supported tools

Manage file permissions when initializing logger

Sometimes the user doesn't have permission to create a file

TMT: remove all NA PSMs before further operations

refactor .aggregateMonoisotopicPeaks

Add Philosopher converter

Final decision on the class structure

Re-write model-fitting functions

• rlm (in feature selection)
• survreg (in imputation)
• lm (in linear summarization)
• medpolish (in TMP)
  all could use a re-write:
• skip unnecessary checks
• remove unnecessary warnings
• return result + information about convergence instead of warnings

DT vs df

Decide, what should be the class of the output (some operations on df may not work with DT) [backward-compatibility]

Filtering intensity

In Skyline converter, missing intensities were only filtered in DDA case:

input <- input[!is.na(input$Intensity), ]

(line 183 in the original file)

Check if the same happened in other converters, be consistent

validObject methods for all classes

Convert .makeAnnotation to a non-internal function

as.data.frame method for MSstatsValidated class

OpenMSTMT: remove duplicated rows

there are identifcal rows in the input that pass all the tests in .aggregatePSMstoPeptideIons -> output has duplicated rows

Same set of parameters for each converter

Consistent names