TPP:4.2.0 Release Notes
From SPCTools
Revision as of 22:12, 5 February 2009 Bpratt (Talk | contribs) ← Previous diff |
Current revision Ntasman (Talk | contribs) |
||
Line 1: | Line 1: | ||
- | == TPP 4.2.0 Release Notes == | ||
- | |||
To developers: | To developers: | ||
Please add anything you want to be sent out here. I'll take out the author attribution in the actual announcement. We can keep this page (or something similar) on the Wiki, as well. | Please add anything you want to be sent out here. I'll take out the author attribution in the actual announcement. We can keep this page (or something similar) on the Wiki, as well. | ||
- | (From Natalie) | + | == Important changes to ProteinProphet behavior == |
- | *To list-- various pepXMLViewer improvements/fixes | + | |
+ | Major changes to protein grouping and subsumed protein behavior have | ||
+ | been made, in order to more accurately compute and display these | ||
+ | relationships. Previously, subsumed proteins would sometimes | ||
+ | appear independently of their subsuming protein. Now, the | ||
+ | subsumed protein will appear as an entry in a protein group around the | ||
+ | parent protein. Thus, the meaning of "protein group" has changed: | ||
+ | while there is still a group probability, it is no longer the case | ||
+ | that there is necessarily only one "true" protein in a group. Also, | ||
+ | many more proteins will appear in groups because any network of | ||
+ | proteins with shared peptides, even contributing peptides, are now | ||
+ | counted as a "group". | ||
+ | |||
+ | Two proteins that share a peptide, and yet have | ||
+ | other independent evidence, will not appear in a group together. | ||
+ | However, if a third protein appears with this peptide, they will all | ||
+ | be grouped. | ||
+ | |||
+ | ==from Natalie== | ||
+ | *build system: | ||
+ | **C++ makefile/build: includes are now normalized and relative to src dir. | ||
+ | *pepXMLViewer: | ||
+ | **now validates as xhtml 1.0 | ||
+ | **compatibality with Google Chrome brower | ||
+ | **more resilient error handling | ||
+ | **xpress decimal ratio displayed (sortable) | ||
+ | **xpress min. area filter | ||
+ | **SpectraST links fixed | ||
+ | **filtering on NTT | ||
+ | **choice of L:H or H:L ratios for both xpress and ASAPRatio | ||
+ | *Windows native installer: | ||
+ | **SSRCalc env variable is now properly set (and uninstalled) | ||
+ | *schemas: | ||
+ | **protXML schema updated to revision 6; adds optional calculated neutural peptide mass to indistinguishable peptide element | ||
+ | **pepXML schema updated to revision 13; adds charge states +6 and +7 | ||
+ | *perl | ||
+ | **modifications for using (optional) non-system perl interpreter | ||
+ | |||
+ | ==from Brian== | ||
+ | |||
+ | *ASAPRatioPeptideParser speed improvement from 10x for small compute-bound files to 300+x for larger disk bound files. | ||
+ | *compressed file I/O: | ||
+ | **Gzipped (.gz) pepXML, protXML and mzXML or mzML files are now treated as native formats, for reduced disk usage and easier sharing across networks: | ||
+ | **The various mz(X)ML converters now accept a -g (and --gzip, in many cases) argument to produce .mz(X)ML.gz directly | ||
+ | **tpp_gui.pl allows users to select gzip output from converters | ||
+ | **tpp_gui_config.pl allows for making gzipped mzXML/mzML/pepXML/protXML the default | ||
+ | **You can simply add .gz to your output filename and the pipeline will know to gzip the pepXML and protXML files it produces. | ||
+ | |||
+ | ==from David== | ||
+ | |||
+ | *ASAPRatio: | ||
+ | **N15 labelling support in ASAPRatio | ||
+ | *InteractParser: | ||
+ | **minimum peptide length parameters sets is_rejected attribute for peptide entries that are too short | ||
+ | *PeptideProphet: | ||
+ | **Automatically disable RT model when there is insufficient data or poor correlation. | ||
+ | **Tuning semi-parametric model to improve modeling performance (and reduce FDR) | ||
+ | **Extending MAX_CHARGE to 7 (common in ETD data), charge must be known first | ||
+ | **Auto-detect refinement in XTandem and disable NTT model | ||
+ | **Use is_rejected attribute in pepXML to invalidate peptides of insufficient length. | ||
+ | **Allow IUPAC ambiguous amino acid symbol B to be considered valid for a glyco motif. | ||
+ | **Automatically disable the use_decoy_ flag when no decoys exist with given label | ||
+ | *ProteinProphet: | ||
+ | **New features: Minimum independence parameter, confidence calculation using States et. al.-like algorithm with a Poisson model and some additional controls to tweak MU ( enabled when PROTLEN option is used) , for IPROPHET print alternative peptide forms as indistinguishable_peptides in the protXML file. | ||
+ | **Improved viewing of iProphet results now displays alternative peptide version (modifications, charges) | ||
+ | *iProphet: | ||
+ | **Adding back true-NSE statistic | ||
+ | **using unified EM approach for all models | ||
- | (From Brian) | + | ==from Luis== |
- | ASAPRatioPeptideParser speed improvement from 10x for small compute-bound files to 300+x for larger disk bound files. | + | *Petunia: |
+ | **New iProphet tab, plus support within xinteract | ||
+ | **Added NOMNC option to xinteract | ||
+ | **Fix link to Tandem output file when input is mzML | ||
+ | **Removed a couple of redundant 'use' statements | ||
- | Gzipped (.gz) pepXML, protXML and mzXML or mzML files are now treated as native formats, for reduced disk usage and easier sharing across networks: | + | ==known issues== |
- | The various mz(X)ML converters now accept a -g (and --gzip, in many cases) argument to produce .mz(X)ML.gz directly | + | *mzWiff: charge state information is not recorded in output mzXML |
- | tpp_gui.pl allows users to select gzip output from converters | + | *massWolf: charge state information is not recorded in output mzXML |
- | tpp_gui_config.pl allows for making gzipped mzXML/mzML/pepXML/protXML the default | + | *trapper: if dual-mode file is processed, the resulting file is possibly always converted as centroid (even if centroid is not selected.) |
Current revision
To developers: Please add anything you want to be sent out here. I'll take out the author attribution in the actual announcement. We can keep this page (or something similar) on the Wiki, as well.
Contents |
Important changes to ProteinProphet behavior
Major changes to protein grouping and subsumed protein behavior have been made, in order to more accurately compute and display these relationships. Previously, subsumed proteins would sometimes appear independently of their subsuming protein. Now, the subsumed protein will appear as an entry in a protein group around the parent protein. Thus, the meaning of "protein group" has changed: while there is still a group probability, it is no longer the case that there is necessarily only one "true" protein in a group. Also, many more proteins will appear in groups because any network of proteins with shared peptides, even contributing peptides, are now counted as a "group".
Two proteins that share a peptide, and yet have other independent evidence, will not appear in a group together. However, if a third protein appears with this peptide, they will all be grouped.
from Natalie
- build system:
- C++ makefile/build: includes are now normalized and relative to src dir.
- pepXMLViewer:
- now validates as xhtml 1.0
- compatibality with Google Chrome brower
- more resilient error handling
- xpress decimal ratio displayed (sortable)
- xpress min. area filter
- SpectraST links fixed
- filtering on NTT
- choice of L:H or H:L ratios for both xpress and ASAPRatio
- Windows native installer:
- SSRCalc env variable is now properly set (and uninstalled)
- schemas:
- protXML schema updated to revision 6; adds optional calculated neutural peptide mass to indistinguishable peptide element
- pepXML schema updated to revision 13; adds charge states +6 and +7
- perl
- modifications for using (optional) non-system perl interpreter
from Brian
- ASAPRatioPeptideParser speed improvement from 10x for small compute-bound files to 300+x for larger disk bound files.
- compressed file I/O:
- Gzipped (.gz) pepXML, protXML and mzXML or mzML files are now treated as native formats, for reduced disk usage and easier sharing across networks:
- The various mz(X)ML converters now accept a -g (and --gzip, in many cases) argument to produce .mz(X)ML.gz directly
- tpp_gui.pl allows users to select gzip output from converters
- tpp_gui_config.pl allows for making gzipped mzXML/mzML/pepXML/protXML the default
- You can simply add .gz to your output filename and the pipeline will know to gzip the pepXML and protXML files it produces.
from David
- ASAPRatio:
- N15 labelling support in ASAPRatio
- InteractParser:
- minimum peptide length parameters sets is_rejected attribute for peptide entries that are too short
- PeptideProphet:
- Automatically disable RT model when there is insufficient data or poor correlation.
- Tuning semi-parametric model to improve modeling performance (and reduce FDR)
- Extending MAX_CHARGE to 7 (common in ETD data), charge must be known first
- Auto-detect refinement in XTandem and disable NTT model
- Use is_rejected attribute in pepXML to invalidate peptides of insufficient length.
- Allow IUPAC ambiguous amino acid symbol B to be considered valid for a glyco motif.
- Automatically disable the use_decoy_ flag when no decoys exist with given label
- ProteinProphet:
- New features: Minimum independence parameter, confidence calculation using States et. al.-like algorithm with a Poisson model and some additional controls to tweak MU ( enabled when PROTLEN option is used) , for IPROPHET print alternative peptide forms as indistinguishable_peptides in the protXML file.
- Improved viewing of iProphet results now displays alternative peptide version (modifications, charges)
- iProphet:
- Adding back true-NSE statistic
- using unified EM approach for all models
from Luis
- Petunia:
- New iProphet tab, plus support within xinteract
- Added NOMNC option to xinteract
- Fix link to Tandem output file when input is mzML
- Removed a couple of redundant 'use' statements
known issues
- mzWiff: charge state information is not recorded in output mzXML
- massWolf: charge state information is not recorded in output mzXML
- trapper: if dual-mode file is processed, the resulting file is possibly always converted as centroid (even if centroid is not selected.)