[Biococoa-dev] BCSequence

[Koen van der Drift]

On Aug 24, 2004, at 9:47 PM, John Timmer wrote:

>
>> It's better if we make the method as follows:
>>
>>   - (float) molecularWeight (int) mode
>>
> That's a great idea.  The main reason I was putting in a new method 
> was that
> I added a separate weight for the ambiguous bases that's the average of
> every base they represent, since that would be more accurate than 
> returning
> 0 for estimating the weight.

I think for the time being that will do. But I'll check some of the 
literature to see how this is solved when people do MS on nucleotide 
sequences.

> Yeah, I might want to handle providing options for whether there's a 5'
> phosphate and such.  I'll have to think about that.  Even so, each of 
> the
> subclasses will probably need to override this method because of these 
> sorts
> of issues.

We should think about a class that stores modifications on nucleotides, 
amino acids, etc. I think we discussed this a few weeks ago 
(modifications, I mean). This should be handled the same for all 
BCSymbol classes. A phosphate or methyl group is the same for every 
sequence.

> I'll see if anything occurs to me, but I think we're getting close to 
> moving
> on to either the wrappers for the sequences that hold features and 
> such or
> some of the tools for translation and calculations.

I think we should also discuss on how to make the original BioCocoa 
classes (I/O) compatible with the new classes. Do we maintain backward 
compatibility, or do we start from scratch?

> My ISP and bioinformatics.org are not speaking to each other today, so 
> I'm
> done committing for today.   I'm going to work on a small program that 
> links
> to the framework and lets you do sequence transformations. To test 
> whether
> everything is working.


Sounds good - we can add it to the project in a folder 'demos'


- Koen.