[Biococoa-dev] more ramblings

Alexander Griekspoor mek at mekentosj.com
Thu Dec 2 02:16:57 EST 2004

>> I think we should choose here, at the moment we did for 
>> BCSequenceType. Or, we adopt the whole BioJava approach with general 
>> sequences, "typed" bases on the used BCSymbolSet. But I don't see a 
>> real mix. In addition, we had many discussion about the BioJava 
>> approach and I still see no clear advantage, even more there are some 
>> real problems associated with it.
> When using the BCSequenceType, we just use plain DNA, RNA, and 
> protein. There is no info about the symbols that belong to it, eg is 
> it  'strict', 'ambiguous', etc. Using symbolsets, we actually define 
> the possible symbols. And we could also differentiate easily between:
> [BCSequenceDNA dnaSequenceWithString: string skippingNonBases: NO];
> and
> [BCSequenceDNA dnaSequenceWithString: string skippingNonBases: YES];
> and even extend the possibilities.

That makes sense, indeed. Also in the current setup we can have three 
methods (or even extend it more):

[BCSequenceDNA dnaSequenceWithString: string];
[BCSequenceDNA dnaSequenceWithString: string skippingNonBases: YES];
[BCSequenceDNA dnaSequenceWithString: string skippingNonBases: YES 
usingSymbolSet: [BCSymbolSet dnaStrictSymbolSet]];

The first two are convenience methods for the last "mother method".

One point Koen makes is that the symbolset preserves whether the 
symbolset used is strict or not, but we could do this as well by adding 
      the BCDNAStrictSequence, BCRNAStrictSequence, 
BCProteinStrictSequence types, which I think is very nice anyway. We 
could implement this even in our sequence-guessing methods.

The great advantage on having typed sequences as subclasses is that it 
isn't complex to implement simple DNA- or protein-only methods in 
one-liners/convenience methods (requiring shared objects, limiting the 
number of tool objects etc). Now I understand why you want to limit 
those to the superclass only, that would fit more in this concept. But 
again, let's choose for typed subclasses OR symbolset typed sequences. 
And if we do, make the greatest use of them in allowing them to 
differentiate in the subclass stage (hey, why would we do it anyway if 
we keep them similar).

                       ** Alexander Griekspoor **
                 The Netherlands Cancer Institute
                 Department of Tumorbiology (H4)
           Plesmanlaan 121, 1066 CX, Amsterdam
                     Tel:  + 31 20 - 512 2023
                     Fax:  + 31 20 - 512 2029
                    AIM: mekentosj at mac.com
                     E-mail: a.griekspoor at nki.nl
                 Web: http://www.mekentosj.com

           LabAssistant - Get your life organized!

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 2945 bytes
Desc: not available
URL: <http://www.bioinformatics.org/pipermail/biococoa-dev/attachments/20041202/8010d706/attachment.bin>

More information about the Biococoa-dev mailing list