[StPeter] - Implemented the changes David requested, rolling back the code to the last stable version and creating the new features from scratch. Use the -r option to enable separate by run analysis.
[StPeter] - rolling back David's changes because it destroyed the correct implementation for many existing datasets. Rather than parameterize the changes necessary to analyze David's one dataset, the entire functionality of StPeter was changed so that all datasets must be analyzed as if they were acquired identical to that one dataset. This broke the fundamental functionality of StPeter, and produces incorrect analyses for thousands of other datasets, including the entirety of the LC data col...
[StPeter] - bugfix when processing n-terminal modifications.
[mzParser] - remove scan reading interrupt when reading scan headers.
[mzParser] : silence error message when precursor scans can not be found in the index because they've been excised from the file.
[mzParser] fix spectrum flags to know the difference between having an ion mobility associated with the spectrum and having an actual ion mobility data array.
[mzParser] - more restoring of code lost in the previous merge. NOTE: there is additional code that has still not been restored: this is deliberate to remove cruft.
[mzParser] - sorry everyone, lost a few lines in a bad merge on my part. Restoring them here.
[mzParser] - Major Change #1: Added ability to read mzMLb file format. As with mz5 format, HDF is required. To compile with HDF support, use MZP_HDF compile time flag (replaces MZP_MZ5), which will include support for all HDF formats (MZ5 and mzMLb). Note that mzMLb.gz is not yet supported, but can probably be folded in with little effort.
[tpp2mzid] More bugfixes related to correctly identifying peptide evidence from ProtXML peptide annotations (and their various indistinguishable modifications).
[NeoPepXMLParser] improved parsing for a few elements. Shouldn't throw warnings when running anymore.
[tpp2mzid] - fix bug when converting protXML protein groups containing peptides with many indistinguishable I/L sequence swaps.
[mzParser] - fix buffer overrun when file name sizes are really, really, really, large. Also some basic code cleanup and recasting to silence warnings that appear on more sensitive compilers.
[mzParser] Fix issues with parsing correct precursor ion from higher msLevel (>2) spectra.
[Comet] - Applying patch to fix bug with combined C-terminal protein modifications and variable modifications. Patch courtesy of Jimmy Eng, ahead of official Comet release schedule.
[Kojak] Updating the README file.
[Kojak] Updating to version 2.0.3
[mzParser] - report error to user when author of an mzML file leaves bugs in their indexes.
[mzParser]: maintaining the sanctity of case-sensitivity in XML
[Kojak] - Upgrading to version 2.0.0alpha22. See kojak-ms.org for complete details.
[tpp2mzid] - fixed bug where SpectrumIdentificationItem id numbers were not incrementing properly.
[NeoPepXMLParser] - Forgot to update the makefile for building this library. Sorry everyone!
[tpp2mzid] - Two fixes: updated the NeoPepXMLParser library to process the <lability> elements that have appeared in pepXML. Also added a new parameter, -d, to specify a decoy prefix string for decoy sequences in the FASTA search database. The default value is "DECOY", but the parameter is used to specify any prefix for decoy sequences, which will then be flagged as isDecoy="true" in <PeptideEvidence> elements of the resulting mzID files.
[NeoPepXMLParser] - Added new element <lability> to <ptmprophet_result>. I have no idea what this is because it is not in the PepXML schema. Also updated some other elements that seem deviant from the schema.
[Magnum] - updated to version 1.0.0 alpha 7
[tpp2mzid] - fixed bug in handling peptide elements without indistinguishable peptide elements. Updating the version tracking to include identification of NeoProtXMLParser library version.
[libraries.mk] - updating build instructions for NeoProtXMLParser
[NeoPepXMLParser] - updating the internal version identification
[NeoProtXMLParser] - Adding new elements to the API for handling Libra-quantified Proteins. Also updated the API for handling peptide elements.
[NeoPepXMLParser] - Fixed a typo when parsing CV and fixed N-terminal mod parsing.
[SPACEPro] - Improved quant starting point to be the PSM with the highest TIC when multiple PSMs map to the same precursor signal. Also added quant support for multiple FAIMS CVs.
[Magnum] - Update to 1.0.0. alpha 6
[Libra] - Fixed SPS/MS3-based quant analysis in LibraPeptideParser. Briefly, when an identified MS2 scan does not have a paired MS3 scan, DO NOT default back to quantifying the peptide from the MS2 scan. This behavior is guaranteed to fail when using, for example, TMTpro16 label with ion trap MS2 scans, because such scans cannot resolve the TMT channels. Instead, those scans simply report no quantitation.
[mzIMLTools] : Update modification table for mTRAQ8plex
[tpp2mzid] : Adds modifications from PTMProphet parameters to the SpectrumIdentificationProtocol for a database search.
[mzIMLTools] - Added formylation to the modification table. Updated modification param addition to be a bit smarter, and to use fuzzy math for different user precisions on modification masses.
[tpp2mzid] Fixed conversion issues with some PTMProphet modification annotations. Added version identifier for the NeoPepXMLParser library that is under the hood.
[mzIMLTools] various speed improvements.
[Makefile] Updating libraries.mk to include latest NeoPepXMLParser elements.
[NeoPepXMLParser] Adding new elements to parse
[mzParser] - Fixed duplication of precursor elements introduced during r8143. Increased additional precursor buffer to accommodate 10 or more precursor ions per spectrum, such as during SPS collection of MS3 spectra.
[mzIMLTools] - added some new shortcut functions and updated adding database behavior to better detect duplicate entries with alternate paths.
Fixed a bug with the --shave parameter.
Updating makefile to build new classes in NeoMzMLParser.
Added classes for Sample and SampleList elements. Updated reading behavior to detect and correct (with a warning) errors in the mzML being read. In this case, specific observations seen in the wild such as not specifying software version and having ScanLists with no Scans.
[mzIMLTools] - Adding additional classes to the API
[mzIMLTools] - updating the API with various fixes used during creation/conversion to mzID
updating a makefile
[tpp2mzid] - bugfixes and improvements for PTMProphet to mzID.
[tpp2mzid] - Interim update for the purposes of record-keeping. Otherwise, my forthcoming changes will be agonizing to undo if needed.
[NeoPepXMLParser] - Updating API, mostly to better support indentation in exported pepXML. Not yet a complete process.
[NeoPepXMLParser] - Updating the PepXML API in anticipation of new features for tpp2mzid; just waiting on Eric for some sample files...c'mon Eric!
[NeoPepXMLParser] - updating makefile for the library
[tpp2mzid] - Updating software to work with the newer mzIMLTools API.
[mzIMLTools] - Adding all the API objects that I forgot that I had written. Oops.
[mzIMLTools] Upadating API
[MikesMagicalMzMLShrinker] - Added some verbosity to the user and also export the params to the meta information in the shrunken mzML. This helps improve clarity and accountability, especially if users start shrinking their magically shrunken files (shrunken-squared).
[common.mk] - updating common.mk to compile all the crap I just added to the TPP.
[MikesMagicalMzMLShrinker] - new tool to shrink mzML files. Its most basic functions remove white space and truncate user-defined identifier tags to a minimal size. More advanced features adds MS-Numpress compression to the files. No matter what, the data are zlib compressed (if they were not zlib compressed previously). Only works with mzML files. Not for mzXML. Not for mzML.gz. However, for mzML.gz, you can gunzip it, magically shrink it, then gzip it back up.
[libraries.mk] - Updating the library makefile to build the NeoMzMLParser API library: libneomzml
[NeoMzMLParser] - a new, object API for reading and writing mzML files. Also supports scan-by-scan iterative reading or writing. Still haven't figured out a good way to reproduce new SHA-1 hashes. Does anyone care about those? I've never actually checked the SHA-1 in an mzML file.
[Comet] - quick little patch that allows MS-Numpress PIC data to be analyzed. The ability was always there, but it was unplugged. It is now active. Also, the next official release of Comet will have this fix in it.
[mzParser] adding const to a string parameter
Quick fix to the MSToolkit code to process MS-Numpress Pic data - it was currently ignoring this.
[MZParser] - recognize all types of MSnumpress+zlib binary data arrays.
[SPACEPro] - Renaming EPICdigest to SPACEPro. Yep, more name changes than Chad Ochocinco.
[EPICdigest] - still more instances of DEEPsearch to scuff up and replace.
[EPICdigest] - diverting the GUI from DEEPsearch to EPICdigest
[EPICdigest] - more makefile updates. Sorry for the spam.
[EPICdigest] - version tracking and updating Makefile
[EPICdigest] - more renaming.
[EPICdigest] - Renaming DEEPsearch to EPICdigest. I expect this will process will go poorly for me. You shall all bear witness to my struggles.
[DEEPsearch] - fixing a bug in specificity calculation that I introduced when simplifying the user parameters.
[DEEPsearch] - Can't make the commandline intelligent, but perhaps I can make it less stupid.
[DEEPsearch] - Made the default no-cut amino acid nothing. User must specify if, for example, they don't cut at proline or any other amino acids.
[DEEPsearch] - Automatically check XML file for use of PeptideProphet or iProphet. Removed command line parameter requiring user input for this. Added new command line parameter and code to allow for n-sense enzymes.
[DEEPsearch] - fixed bug where modification was on enzyme cut site. Also updated JSON output to have a .ds.json extension to specify it originated from DEEPsearch.
[DEEPsearch] : Adding new tool to quantify and evaluate digestion efficiency at the PSM, peptide, and protein levels.
[TPP_GUI] Mike can't type.
[StPeter] [tpp_gui] - adding additional parameter to StPeter launch page of the tpp_gui
Small change to squeeze a little more speed (my test show 5% improvement, meh...) from the DecoyXcorr operations.
Changes to fill method for SpScore sparse matrix. Use less computation time, but a tad more memory. Still much more memory efficient than full matrix.
Created sparse matrix structure and changed code to use it. Basic operation is as follows:
Fixed two bugs requesting values outside array boundaries. Fixed uninitialized values to stop valgrind complaints.
Removed iterative allocation/deallocation of temporary arrays in the CometPreprocess::Preprocess() function. Memory is now allocated once prior to the first CometPreprocess::Preprocess() call, and deallocated after the last CometPreprocess::Preprocess() call. The temporary arrays are double dimension, one for each thread, and a lightweight management system is used to indicate when each thread is done with the temporary memory arrays.
Extended single allocation/deallocation of preprocessing arrays to include CometPreprocess::Smooth() and CometPreprocess::PeakExtract(). Arrays are allocated once, using a multi-dimensional array equal to the number of threads being used.
Added a basic, thread safe profiler for tracking time spent in any function. Not sure what the units are. Sure as hell not seconds. Left in a few macros to do the conversion if you are so inclined.
Overhauled the sparse matrix functionality to improve performance. The old sparse matrix could add significant overhead when used improperly or when parameters or data were not optimal. The sparse matrix structure was reduced to a two-dimensional array of floats where one dimension can be left uninitialized as needed (i.e. when it would only be filled with 0 values) to preserve memory. Values are grabbed from the array with at best a single operation (division) to convert the requested bin, a...
Fixes a bug identified and solved with the help of Patrick Pedrioli. Briefly, when applying large precursor mass tolerances (as per recent Gygi lab publication), Comet would crash. The bug was traced to the spectra arrays that apply a cushion of 100 Da over the precursor mass. When searching with tolerances greater than 100 Da, peptides would appear whose fragment ions exceeded the boundaries of the spectra arrays.
Undoing accidental upgrade to a variable used in the newest MSToolkit. Comet has not yet rolled in this newest MSToolkit, and thus this change should wait.
Updated CometWritePepXML to use .mzML.gz and .mzXML.gz for the msms_run_summary raw_data attribute instead of .gz for gzipped mz(X)ML files.
[NeoProtXMLParser] - a new C++ API for reading and writing ProtXML files. This is intended to *eventually* replace the libprotxml currently in the TPP.
[NeoPepXMLParser] - a new C++ API for reading and writing PepXML files. This is intended to *eventually* replace the libpepxml currently in the TPP.
[RACI]: potentially fixing potential memory leak.
[StPeter2Matrix] - Nearly forgot to add this in advance of its TPP workshop (oops!). New utility for combining multiple
[DidIScanThat] : Added new option to additionally export results to JSON file. Added more cerr messages.
[comet] Put a patch for mzML.gz files in pepXML export. This patch won't be necessary for the next public release of Comet.
[fixPepXMLScan] : Added new tool that can update (to a new file - no overwriting) pepXML files that
[mzParser] : Added some functions to access index data. Updated some function parameters for improved compatibility.
[TPP] : Updated makefile to no longer compile boost and pwiz when building SpectraST, because SpectraST doesn't use them. Nifty!