Activity for Comet MS/MS search engine
-
keesh posted a comment on ticket #28
Jimmy, We employ Comet in a clinical Proteomics pipeline at Mayo. We would like deterministic results to adhere to great patient quality. We would be happy to submit a pull request in order to facilitate this, if you are too busy. Can you provide some hints on the affected variables / threads?
-
Jimmy Eng committed [r1625]
update to GA4 tag
-
fix link to percolator.ms
-
addressed in branch: https://sourceforge.net/p/comet-ms/code/HEAD/tree/branches/release_2020010_cyclic/
-
support searching cyclic peptides
-
Search light/heavy SILAC as mixed spectrum
-
addressed in branch: https://sourceforge.net/p/comet-ms/code/HEAD/tree/branches/release_2019015_silacpair/
-
allow specifying two sequence databases for a search
-
addressed with GH commit 75d837
-
Annotate decoy sequences in Percolator pin output when not using Comet's internal decoys
-
clip N-term methionine and PEFF
-
addressed with git commit 450b6a
-
Add mzIndentML output support
-
Initial mzIdentML support in release 2020.01.0 released 2020/11/09
-
allow specifying two sequence databases for a search
-
update website content
-
propagate r1620 fix to SILAC pair branch so that paired peaks can be scored even if base fragment ion peak is not matched.
-
simple change to the version string for this variant of the tool
-
fix to allow add_fragment_masses to match when base fragment ion is not present
-
add_fragment_masses should now be functional with these updates
-
Initial code changes for applying user defined fragment mass additions/offsets via the "add_fragment_masses" parameter. Extends the 4th field of uiBinnedIonMasses to allow up to 9 fragment mass additions. Not supported for indexed searches.
-
add image
-
delete unused branch
-
create iqPIR branch from silacpair code
-
delete; wrong location
-
website files
-
create iqPIR branch from silacpair code
-
Add in code to calculate xcorr & E-values for base peptide fragments and paired only peptide fragments.
-
Grab the SILAC pair mass difference from the first variable mod entry with a lysine residue. This replaces the hardcoded 8.014199.
-
Comet MS/MS search engine updated /comet_2021010.zip
-
-
fix the total ions calculation
-
add lnExpectPair column to Percolator pin output
-
track and store xcorr and E-value for contributions from just the paired fragments
-
little mstoolkit readme update
-
use an actual correct variable name in total ion count calculation
-
mzIdentML: change "DBSequence_ref" to "dBSequence_ref" thx to note by A. Collins.
-
Add "cyclic_NL_ms_level" parameter to control applying cyclic neutral loss to MS2, MS3, or both. Add support to consider/analyze higher charged fragments. Calculate total ions correctly for linear vs. cyclic peptides. Still need to address Sp score, matched ions, and E-value calculation.
-
Correct the # of reported total ions. Previously the number was calculated based on fragment charge states through (precursor_charge - 1) but this did not take into account the max_fragment_charge parameter entry whose value is often less than (precursor_charge - 1).
-
Issue is current implementation checks szProteinSeq (which has start methionine clipped off) against the original dbe->strSeq within StorePeptide(). With a PEFF, szProteinSeq will now also contain a substitution so the strcmp() fails. Need to do a strcmp after substituting back in the original sequence unless I implement a more direct way around this round-about check like simply setting a new variable.
-
merge from branches/release_2021010 encapsulating r1586 thru r1598
-
Issue is current implementation checks szProteinSeq (which has start methionine clipped off) against the original dbe->strSeq. With a PEFF, szProteinSeq will now also contain a substitution so the strcmp() fails. Need to do a strcmp after substituting back in the original sequence unless I implement a more direct way around this round-about check like simply setting a new variable.
-
-
tagging release 2021.01 rev. 0
-
delete tagged release to fix sequence pre-load error
-
remove errant comment on line limiting pre-loading to 500 sequences
-
-
tagging release 2021.01 rev. 0
-
delete 2021.01.0 tagged release
-
add a cast to allow VS2019 win32 compile thx to fix by D. Shteynberg
-
tagging release 2021.01 rev. 0
-
cleanup unused declarations in header file
-
bandaid workaround until there's a fix for precuror NL peaks bug where _uiBinnedPrecursorNL values are < 0
-
remove print line
-
Add in safety check and error message in XcorrScore() prompted by large negative "bin" value when "precursor_NL_ions = 1" is set.
-
In XcorrScore, the precursorNL "bin" has intermittently been observed to be a large negative value; haven't tracked down why though.
-
add no digestion entry to comet.params
-
update the README to document using v142 build tools in VS2019
-
update CometUI.pfx temporary key
-
since no parameters change, allow the use of 2020.01 params files also
-
create 2021.01.0 branch
-
Not sure if I can replicate this anymore.
-
add preliminary 2021.01.0 release docs; nothing updated yet
-
Extend protein sorting function ProteinEntryCmp() to consider peptide start position in order to get the first set of flanking residues when a peptide is present in a protein multiple times.
-
In CalculateSP(), evaluate code below. double dNewMass = dBion + g_staticParams.massUtility.pdAAMassFragment[(int)pOutput[i].szPeptide[iLenMinus1]] + (pOutput[i].piVarModSites[iLenMinus1] != 0 ? pOutput[i].pdVarModSites[iLenMinus1] : 0.0) + g_staticParams.staticModifications.dAddCterminusProtein + Oxygen_Mono + Hydrogen_Mono + Hydrogen_Mono;
-
investigate ThreadPool
-
David implemented new ThreadPool beginning with r1541
-
David implemented new ThreadPool.
-
add "old_mods_encoding" parameter entry to write modified peptide strings using the old mod character encodings for .sqt output
-
update version string; use wait_on_threads() instead of sleep() in CometSearch()
-
merge from branches/20210524-ThreadPoolUpdate
-
usleep() is deprecated; use nanosleep()
-
remove CheckForUpdates from Makefile; use portable sleep
-
given new ThreadPool implementation, add logic to not load all protein sequences at once to better match previous implementation behavior
-
remove unused variable
-
minor change of performance debugging reporting
-
David Shteynberg committed [r1574]
Have CometSearchManager allocate the ThreadPool only once per execution of comet...
-
Remove thread pool from DoSingleSpectrumSearch() which entails overloading RunSearch().
-
Update to Windows line terminations.
-
Code clean up; remove unused min/max threads arguments from RunSearch(), PostAnalysis(), etc.
-
Revert for loop range for E-value calculation from r1555 check in.
-
fix formatting to conform to project coding style
-
Skip thread pool initialization for DoSingleSpectrumSearch(). Upgrade projects to VS2019 v142.
-
fix logic for y-ion silac fragment pairs only
-
add option to use y-ion silac fragment pairs only by setting "silac_pair_fragments = 2"
-
The spectrum batch loading was exiting prematurely due to checking total scans analyze vs. spectrum batch size; removing this check restores proper behavior.
-
revert the for-loop bounds change
-
remove histograms from text output and fix for-loop bounds in E-value calculation
-
remove return value for function of type void
-
Search progress percentage should be updated when the thread finishes, so the percentage completed is meaningful ...
-
New ThreadPool uses exact number of threads, the main thread also runs the jobs, tested with 1 to maximum threads, appears fractionally faster than trunk code and will finish even of giant mzML files like those that come out of timsTOF DISCo output ...
-
replace usleep() and sched_yield() with standard counterparts
-
add a wait to the spectra loading to limit the number of spectra loaded to a number near the spectrum_batch_size as with the previous implementation
-
Makefile fix
-
add "-std=c++11" to Makefiles to allow linux compile
-
address the new E-value calculation inconsistency across replicate runs
-
remove -g compile option from MSToolkit's Makefile
-
revert trunk to r1543 prior to ThreadPool changes
-
create branch for David's ThreadPool update
