Re: [Seqan-dev] HMM Demo with modified alphabet

[Manuel Holtgrewe <manuel.holtgrewe@fu-berlin.de>]

Hi Konstantin (? or Klaus or Mark?)

You might want to reconsider whether you want to use the  
implementation SeqAn. Currently, the HMM specialization of the class  
Graph assumes that the number of states is fixed at compile time and  
changes to this would require major changes in the code and any code  
that uses it.

If you want to use this, however, you can simply define the alphabet  
as a new Simple Type (analogue to TraceBack in module graph_align, for  
example):

struct _MyType;
typedef SimpleType<unsigned char, _MyType> MyType;  // use unsigned  
short for > 255 values

For this type, you then at least have to specialize the ValueSize<>  
and BitsPerValue Metafunctions:

template <> struct ValueSize<MyType> { enum { VALUE = 128 }; };
template <> struct BitsPerValue<MyType> { enum { VALUE = 7 }; };  //  
must be ceil(log(value size))

Then, try to parametrize the HMM graph class with this alphabet and  
add any missing functions, metafunctions and operators for your new  
alphabet. However, deriving your new type from SimpleType, should  
automagically use the code for SimpleType that the DNA and Amino Acid  
alphabets use, too.

Cheers,
Manuel

Am 05.09.2010 um 22:43 schrieb Klaus Klapperon:

> Hello again,
>
> I want to create a hidden markov model with an alphabet of (at  
> least) 128 values (it has nothing to do with bioinformatics). I  
> started with the hmm demo (graph_hmm.cpp). But how can I define this  
> new alphabet type?
> Viele Grüße,
>
>
> Konstantin
>
> --
> GMX DSL SOMMER-SPECIAL: Surf & Phone Flat 16.000 für nur 19,99 Euro/ 
> mtl.!*
> http://portal.gmx.net/de/go/dsl
>
> _______________________________________________
> seqan-dev mailing list
> seqan-dev@lists.fu-berlin.de
> https://lists.fu-berlin.de/listinfo/seqan-dev