Part II

Skip to end of metadata
Go to start of metadata

Content

Deliverables

This showcase has the following deliverables

  • Final version of workflow (test workflow content-wise: validate with data from literature that the workflow performs the analysis it is supposed to do).
  • ROification (using the RO manager and the RO portal)
  • Provenance (using latest provenance export options in Taverna)
  • Outline of a paper for the Journal for Biomedical Semantics paper, SWAT4LS issue. The paper will serve as the LUMC reference paper for wf4ever-related work.

JIRA

WFE-589

Materials and Methods

  1. MyExperiment:
    1. main workflow: http://www.myexperiment.org/workflows/2973
    2. pack: http://www.myexperiment.org/packs/282
  2. RO manager: https://github.com/wf4ever/ro-manager/tree/master/src
  3. RO catalogue on github: https://github.com/wf4ever/ro-catalogue/tree/master/v0.1/concept-profile-matching-golden-exemplar
  4. RO in RO Portal: http://sandbox.wf4ever-project.org/rosrs5/ROs/concept-profile-matching-golden-exemplar/
  5. RO models: http://wf4ever.github.com/ro/
  6. Protege: http://protege.stanford.edu/

Blog

13-07-2012

Skype meeting whole day with Graham and Reinout. Summary (by Graham):

Kristina and Reinout have RO manager installed and running

  1. A copy of Kristina's "anni" exemplar workflow is now in ro-catalog github project
  2. Graham has a copy of anni that can be loaded into Taverna (but I haven't done a complete load yet, or tried to run it)
  3. We have "RO-ified" the anni workflow (minimal RO structure on anni files)
  4. We have figured out how to use RODL OpenID authentication
  5. We have pushed a copy of basic anni RO to RODL
  6. We have added some very simple annotations to the "anni" RO
  7. We have identified some weaknesses in ro-manager annoitation facilities that need improving
  8. We have identified some initial information that we can add to anni to enrich its value as an RO

Plans going forward:

  1. Kristina will give further thought to structuring and enhancing the anni RO based on what we covered today
  2. Graham will improve the ro-manager annotation facilities to better handle links and simple (semantic) annotations
  3. We will return to the best practices principles and start to understand how they can be supported by checklist evaluation

16-07-2012

  • Read the e-science paper submission (Jun et al.) and gave my comments.
  • Read the RO-model related papers in the Dropbox.
  • Skype meeting with Harish. He provided the last version of his workflow to be added to the RO. Have not tested it yet.
  • Updated my workflow with his last version of the SNP2EntrezGene component. Now it maps to more genes, which results in a very unstructured output of my workflow.

17-07-2012

Anni update sprint in Rotterdam (Erasmus MC biosemantics group) with Reinout, Erik van Mulligen (EMC), and Bharat Singh (EMC). Outcome: The document database (Medline) is being updated (will probably finish over night). Went to the code with everyone and made the necessary changes of input variables. Bharat will start the script when the Medline database update has finished. Bharat took notes and edited the Biosemantics wiki.

18-07-2012

  • Restructured my RO so it will be easier to annotate and understand. Removed obsolete and old test files and old versions of workflows. Files are sorted in subfolders. New name: "concept-profile-matching-golden-exemplar"
  • Tried to commit my changes to github. Ran into newbie problems but with the help of Graham I managed in the end. Note to myself when working with git:
    • Start with "git pull" to sync changes made by others
    • Make local changes
    • Then use "git add <filename or directory>" to add files and "git rm <filename or directory>"
    • Use "git status" to check changes lined up
    • Use "git commit" to commit
    • Use "git push" to finalize the changes (need to fill in user name and password)
    • If I have messed up my own folder and want to clone a new copy: "git clone"

20-07-2012

  • Tried to push the changes to RORSRS with the help of Piotr. 
  • After a lot of trial-and-error I found out that "push" ($RO push) works if I have the right ROBASE and RO settings in the "ro.sh.kmh-kmhettne" script in the ro-catalogue and the right configuration settings in the make.sh script in the RO. See github folder for these scripts. 
  • Tried to move the RO annotation as step further than the trivial RO annotation that is created by the RO manager basic script. Wanted to annotate a workflow as a workflow This is how to do it:
    1. Select the resource
    2. Under "Annotations for the selected resource" click "Add annotation"
    3. Fill in the boxes:
      1. What kind of statement is this?: Other
      2. Custom property URI: http://www.w3.org/1999/02/22-rdf-syntax-ns#type
      3. Statement object type: click this box
      4. Resource URI: http://purl.org/wf4ever/wfdesc#Workflow
  • Looking that the wf4ever ontologies however, I am confused as to if I should use Workflow since it seems to refer to an abstract workflow. This is not clear to me. I probably need to talk to Jun and Stian.

21-07-2012

Tied up all day due to a visitor (David Wild from the Indiana University School of Informatics).

22-07-2012

Wrote the outline of the paper.

Tried to figure out how to proceed with the model mapping.

Comments:

  • To see my changes I had to click on MY RESEARCH OBJECTS and follow the link for my research object
  • Not sure if/how I can/should annotate the relations?

Summary

A final version of the workflow (and its nested components) were created and validated by Kristina. A pack on MyExperiment contains the main workflow and its nested components (http://www.myexperiment.org/packs/282). A first version of the RO was created using the wf4ever tooling with a lot of help from Reinout, Graham and Piotr, and can be found here:

  1. the RO catalogue on github: https://github.com/wf4ever/ro-catalogue/tree/master/v0.1/concept-profile-matching-golden-exemplar
  2. in RO Portal: http://sandbox.wf4ever-project.org/rosrs5/ROs/concept-profile-matching-golden-exemplar/

Kristina made a provenance export for one of the workflows using Janus, just to be able to experiment with wfprov. The export can be found in the RO.

Kristina wrote the outline of the paper for the SWAT4LS special issue in the Journal of Biomedical Semantics and Marco reviewed it. The outline can be found in the dropbox (https://www.dropbox.com/home/Wf4Ever/Publications/Journal_of_Biomedical_Semantics). Direct link:https://www.dropbox.com/s/7kluijczx0ec5vh/article_outline.pdf

Issues raised

  • At the moment the user needs to know how to use the following tools to be able to work with workflow-centric research objects:
  1. Taverna (create workflows)
  2. Web services (check annotations)
  3. Github (structure and version data before ROification)
  4. RO manager (ROification)
  5. RO portal (ROification and visualization)
  6. MyExperiment (packs)
  7. ROSRS (versioning)
  8. RO models (ROification)
  9. Protege (visualize and navigate RO models to make informed choices)

This is doable, but not without quite a lot of technical support!

  • How to deal with provenance export and a user view on provenance?
  • I want to structure my output in a table. This can be done using a bean shell that formats the output. However, if I then change something in my workflow I would need to change the bean shell as well. Would be much easier if I could construct the table on the fly in Taverna. For now I just leave it for what it is.
  • Coordinated information from multiple WF runs (This is related to the output structure. To avoid a big mess I now run it for one snp at a time). QUESTION: How to handle multiple runs?
  • How to work with myExp PACKs at this point? Graham's suggestion is to focus for now on the RO, and expect the myExp updates to handle this as a Pack.

Meetings

Whole day Skype meeting Graham+Kristina+Reinout 13-07-2012

skype-chat-25-07-2012.txt

Issues for off-sprint and follow-up sprints

  • Study Jira guidelines and edit Jira for this showcase where needed (Reinout)
  • Internal review
    • Guidelines for our internal reviewers
      • Idea: follow the PlosOne principle in context of preservation: methods should be sound, understandable, and testable; how new or exciting the results are is not part of the review.
      • Add information around the workflow (especially I/O) to guide the reviewer in what to test.
    • Idea: Harish' suggestion to see if we can reproduce older papers
  • Scenario for Taverna Table support (or try new functionality if it exists)
Enter labels to add to this page:
Please wait 
Looking for a label? Just start typing.