Week 4, 5

-
Greetings!

This blog contains work done in week 4 and 5. It can be noticed that till week 3 I had completed most of the work in the background, updating Matlab Parser was still left, but neither had I pushed my code, nor did I check my code with ample test cases. Not directly working with the codebase is easy, but once the concept is understood and is clear, it is really desirable to check it by adding to the actual codebase and very importantly get suggestions from the mentors.

Week 4 was hectic and was spent on adding all the work to the codebase and solving the Matlab parser issue. Here is a short detail about each issue and its solutions, solved in evaluation 1:

  • Extend ModelPolisher annotation capability (#39, #41):
    • Issue: Add feature in ModelPolisher to add annotations to species, reactions, and geneProducts which do not have Bigg Ids.
    • Solution: For species, reactions, geneProducts already some annotations, we can map the annotation URIs to their BiggIds. 
    • Details: For species and reactions, BiggId is present in bigg_id table of component and reaction table in BiGG database. For geneProducts, BiGG is present in locus_tag table of gene table. See the actual mapping function(class: BiGGDB.java) here.
      For species and reactions whether the Id is BiggId or not could be checked by syntax but for geneProducts the class SMBLPolisher.java adds "G_" to the id. Thus for geneProduct we needed a dedicated function(class: BiGGDB). After checking if the id is not BiGG Id, functions(class: BiGGAnnotation.java)(species, reactions, geneProduct) were implemented to get the BiGG Id from the URI list.
  • Update Matlab Parser (#37):
    • Issue: The current MatlabParser had problems had issues with java update (not compatible with Java 11). Also, the library was no longer maintained.
    • Solution: Update library to HebiRobotics/MFL.
    • Details: Matlab files are parsed in cobra.COBRAparser.java, and the library is used in just this one file. Thus I basically had to update the dependencies and use of library in this file only. A major issue was faced after implementation: some mat files could not be read by the parser library. My mentor, Thomas, helped me by opening this issue(#40) in the library, after it was solved in PR#41.
  • Minor Issues:
    • Error in building: setupDB failed(#36).
    • Improve Documentation(#31).
These were details in short of the work done before 1st evaluation. I would like to thank my mentor a lot for giving me this opportunity to work with them and for helping me throughout the period.

During week 5, I didn't do much work, except for some formatting and finalization of code (yeah that too takes time ;)). Meeting for week 5 was held on June 28 (Friday) in which the following decisions were made:
  • First, solve the issue of containerization of ModelPolisher before the next evaluation. 
  • Add test cases to check the mapping regarding issue #39, #41.
  • Update ModelPolisher to be compatible with openjdk-11.
  • Another issue of removal of objective from SBML if listOfObjective is not present was discussed. This will also be solved before evaluation 2. 
  • Except these, in the next meeting we shall discuss how we can use annotateDB in ModelPolisher.
Thanks a lot for your time. 

Thank You.

Comments

Post a Comment

Popular posts from this blog

Final Report: GSoC 2019

-
This is the final report for project  Extending ModelPolisher to a universal model annotation tool  under Google Summer of Code 2019, organization: NRNB . ModelPolisher is a model annotation tool originally developed for the BiGG Models Knowledgebase. Annotations enhance the reusability and interoperability of biological models. During this project, I extended the model annotation capabilities of ModelPolisher, updated the MatlabParser (reading models in the COBRA matlab format), containerized ModelPolisher using Docker, added functionality to use AnnotateDB for model annotation, and added a feature to produce a model annotation glossary and CombineArchives as outputs. As part of this work, the SQLite version of the ModelPolisher database backend could be removed. Link to G ithub Project:  https://github.com/draeger-lab/ModelPolisher The project proceeded according to the goals and timeline of the project  proposal ,  except for one major change: I comp...
Read more

Week 12

-
Greetings! After completing the code for the production of the glossary file, my next task for this project was to give the user an option to produce the output in form of a Combine Archive which contains the polished model and the glossary file. This feature may look redundant now, but it is very helpful to keep biological models in differents formats in one Combine Archive which also contains a single glossary file which has all the annotations from the polished model. So, for this task, I first added an option `--output-combine` for users in which they can provide a boolean to produce output as a single Combine Archive. After addition of this option, if the user provides value `true` for this option, I use binfalse/CombineArchive  repository to put in the output polished model and the glossary file in a Combine Archive zip. Note that, in this case, any annotated model or glossary rdf file will not be produced. There were multiple issues during the implementation of thi...
Read more

Week 9, 10

-
Greetings! The following blog contains my progress in Week 9, 10. In the last days of the previous week, I had already integrated AnnotateDB in ModelPolisher but it was not yet properly checked. So, this week my major goal was to completely check the integration of AnnotateDB and solve bugs (if any). Also, I had to update the README file to provide proper instructions for users to add annotations from AnnotateDB. Testing AnnotateDB Integration:   I had already tested the to query to get the annotations from ADB using BiGG Id of species and reactions, so here only ModelPolisher was tested for successful extraction of annotations from ADB. Issue Observed : After I had added support for ADB in ModelPolisher, I tried to polish some random models (after temporarily removing code to extract annotations from BiGG DB). I observed that ModelPolisher is not making queries to ADB while the extracting annotations from ADB in class `AnnotateDB`, but rather the queries are being made to Bi...
Read more