Retreats Fall2010 - OpenMS/OpenMS GitHub Wiki

• each tool prints small !HowTo: data type (raw, centroided...), required preprocessing (which tools?), recommended postprocessing?
• algorithm summaries are missing. In a TOPPAS pipeline one never knows where it went wrong. Hence, print a summary statistic (e.g. #mapped peptides, #of discarded proteins due to...)

08:30-12:00:

Feature Finder (Metabolites)

Modify !CentroidedFeaFi to work with multiple averagine distributions, come up with an easy way to specify these (in particular in the context of stable isotopes and the skewed isotope distributions caused by them) (Chris)

Review applicability to recent labeling techniques (e.g., dimethyl labeling, lanthanide chelates) (Chris)

MS to the e (Sandro)

Metabolomics Feature Finder -- presentation of the current status (Erhan"), benchmark against proteomics !FeaFi, compared to XCMS

TOPPAS workflows durchsprechen

Afternoon:

15:00-18:00: File formats

Overview over OpenMS formats. (Matthias)

Review/extend file formats accepted by !FileConverter (DTA? !mzIdentML? !TraML?)

assess !mzIdentML completeness (Pieter Neerincx from ML?) (Mathias)

discuss first implementation of !mzQuantML (before HUPO world meeting in Sydney two weeks after the retreat!)

Discuss integration/use of new formats: !mzIdentML, mzQuantML, !TraML? Who is going to stay in touch with PSI working groups?

Fileformate: Emanuele Alpi sucht Konverter von MASCOT 2.2 zu mzIdentML -- auch solche (Mathias)

Zwischenformate (Mascot,...) in !IDFileConverter einbauen?

FIX untested/broken tools (!SequestAdapter in particular!). ID team Sven, Sandro, Timo divides up responsibilites. Timo takes on Sequest with the help of Sven and Sandro to get acquainted. Compare SEQUEST interface across versions (Berlin vs Zürich [Chris], if different, the adapter might be hard to maintain --> look at Crux)

Review deprecations

Nonlinear retention time alignment?, estimation of false-matching rates in featurelinker?

Dealing with huge (200 GB?) raw data files – where are the bottlenecks, how to resolve that?

FeatureXML files can get quite big due to new "convex" Hulls - extend XML format?

Quality COntrol Tool? Peptizer (Vaudel, Martens) – spectrum id validation

Can we include !ProteoWizard in the distro somehow to make sure we can read formats?

code stability: not very well tested.

• Empty input containers often lead to a crash.
• Almost empty containers might lead to wrong distribtion estimates with not even a warning given

reorder error messages;

FalseDiscoveryRate:
..\OpenMS\source\ANALYSIS\ID\FalseDiscoveryRate.C(148): Meta value 'target_decoy' does not exists, reindex the idXML file with 'PeptideIndexer' first (run-id='XTandem_2010-06-23T12:32:29, rank=1 of 1)!
..\OpenMS\source\APPLICATIONS\TOPP\FalseDiscoveryRate.C(155): FalseDiscoveryRate failed due to missing information (see above).

to

FalseDiscoveryRate:
Meta value 'target_decoy' does not exists, reindex the idXML file with 'PeptideIndexer' first (run-id='XTandem_2010-06-23T12:32:29, rank=1 of 1)! @ ..\OpenMS\source\ANALYSIS\ID\FalseDiscoveryRate.C(148)
FalseDiscoveryRate failed due to missing information (see above). @ ..\OpenMS\source\APPLICATIONS\TOPP\FalseDiscoveryRate.C(155)

• !MacBook Air als SVN+FTP+Docu server, DHCP, WIKI, (Stephan)
• Switches, Ethernet, power cords, etc.

• Towels and bed sheets