MSA skype 20100615

Back to

Participants:

Mark, Scooter, Kyle, Andreas

Report from last week:

  • The questions regarding phylo and forester library were resolved by email

  • Scooter and Andreas refactored the phylo module so forester is now a stand alone jar file

  • All of BioJava modules now compile and Andreas will move a automated build system into production during the next days.

  • Mark committed the code for pairwise alignments (Needleman Wunsch). There were a few complications, hence slight delay from last week.

Questions

How to deal with default values?

  • Shall there be constructors taking many arguments, or should there be no arguent constructors.

Andreas, Kyle: no argument constructors good for serialization and scripting langages

Mark: solution that is “best of both worlds”. The classes have nested classes with default values. Will commit test cases later today then we can take a look and provide feedback by email.

Help for user to learn API

Scooter: FastaHelper class is an example. How should we deal with helper classes? Andreas: The reference documentation is the BioJava Cookbook page on the wiki.

How to treat additional Info

Kyle: what about additional info like secondary structure, solvent accessibility

Scooter: Sequence is the storage container for those extra values.

Mark: trying to split interface from any class implememtation so people who have different styles of scoring or alignments can provide their own implementations and can plug it into the rest of the routine. i.e. replace one piece while leaving the rest in place.

Kyle: A test case would be a MSA where a developer has some PDBs and use the sec. struc assignment info and create a custom scorer that will use this info for the alignment

Outlook for this week:

- Get similarity matrix from pairwise alignments
- Convert similarity matrix to a guide tree…
- Provide feedback on Mark's code for dealing with default values