[Biococoa-dev] Annotation questions

Philipp Seibel biococoa at bioworxx.com
Sun May 29 16:04:56 EDT 2005


Hey Koen,

nice to hear something going on.

> Hi,
>
> Spent some time today on BCSequenceReader to include more  
> annotations. For the SwissProt format (and probably other formats  
> as well), some annotations are in multiple lines. So I modified the  
> - (void) addAnnotation:(NSString *)annotation forKey: (NSString *)  
> key to test for existing keys. In that case I am appending the  
> annotation, instead of replacing it.
>
> Right now, the code in readSwissProt now first creates an dummy  
> sequence, then iterates over all the lines to add th e annotations,  
> and then  needs to add the sequence to the BCSequence object.  
> Currently there is no 'setSequence' method, so I choose to create  
> another BCSequence object with the sequencestring, and then copy  
> the annotations from the dummy object to the real object. However,  
> the setAnnotations method was commented out. Are there any  
> objections if I uncomment that?
>
> Another approach could be to move the addAnnotation from  
> BCAbstractSequence to BCSequenceReader.
>
> Also, we need to think how to implement nested annotations. The  
> easiest would be to make a BCAnnotations object as the value for  
> another BCAnnotation.

The Annotation and Feature stuff has to be designed very carefully.  
Btw i need it as well in my ProteinStructure classes. I think using  
dictionaries and dictionaries in dictionaries are a solution, but  
very difficult to understand for the user of the framework. On the  
other hand it's very flexible and could be easily saved as a .plist.  
So we should consider where we will need it in the future. I think  
features are even more difficult to realize, because they could have  
the same key and have a special order.

Finaly i would suggest to discuss at least for 2 - 3 weeks, what we  
want to do with features and annotations and of course where they  
appear in "real life". We should take the time, because this is a  
very fundamental structure. I would also suggest to look at the  
implementations of BioJava and the others. Btw. is there a BioJava  
user in our group, to tell about pros and cons of their  
implementation. Ok so far from me .....


cheers,

Phil



More information about the Biococoa-dev mailing list