[Biococoa-dev] BCSequence

[Koen van der Drift]

On Aug 23, 2004, at 5:51 PM, John Timmer wrote:

> Okay, I'm about to commit a major update to BCSequence that eliminates 
> the
> sequence string and reworks some of the related methods and 
> initializers.  I
> also took the time to reorganize the methods a bit to make navigating 
> the
> file a bit easier.
>


John,

A few comments.

You put this in the header for the molecularWeight method:

/*!
     @method    - (float) molecularWeight
     @abstract   returns the predicted molecular weight of the sequence
     @discussion calculates the predicted molecular mass of the 
sequence, based on
     * average isotope use.  Subclasses calculate based on the loss of 
atoms (ie -
     * H20 in peptide bond formation) and use averages for symbols that 
represent more
     * than 1 individual symbol.
*/

It's better if we make the method as follows:

  - (float) molecularWeight (int) mode

where in mode we pass a enumerated constant monoisotopic or average. 
You cannot predict in advance if a client needs the monoisotopic or 
average mass.

Also we should always add water for any sequence (only one 
watermolecule is necessary for a whole sequence), so we might as well 
put it in the same method instead of delegating it to the subclasses. 
The masses in the plist are actually residue masses, so without H2O. I 
will make these changes tonight.

You added the method length, but there is alreasy a method 
numberOfSymbols. However, because the naming of the first one is more 
logical, I will remove the numberOfSymbols method as well.

I'll wait for some input before I commit my changes.

Otherwise, a great improvement! I started with a BCSequenceProtein 
class, but see that most of the sequence manipulation is already in 
BCSequence. Not sure right now what else should go in that class. As 
Alex already suggested, additional functionality (pI calculations, 
digests, etc) should be in helper classes, such as BCProteinTools.

- Koen.