[BioRuby] Ruby & R

[Pjotr Prins]

Hi Raoul,

On Fri, Mar 13, 2009 at 11:57:38AM +0100, Raoul JP Bonnal wrote:
> Hi guys,
> I have always worked on sequences and bioruby gave me everything I need  
> to work with that kind of data.
> Now I moved to Illumina gene expression and rtpcr data, I think there is  
> a lack of support in this kind of data. Isn't it ?
> I found that usually this data are handled by R, but which are the  
> pro/cons of this lang ? And as "lang/syntax"... it's not very  
> confortable I think :-(

R is the language of choice for most statisticians. So it is natural
that where statistics count R gets the functionality. R as a language
is not a good bargain. I see it more as as succesfull 'glue' language
of C libraries (rather slow glue too). I think the only place it
shines is with matrix manipulations. And then there are the libraries -
which, as it happens, are mostly written in C.

> Pjotr is providing support for biolib to Ruby.

As these are C libraries I can easily provide Ruby mappings with
biolib. If you point me at specific R functionality you want I can
see if I can map it.

> Could Ruby do it better ?

I don't think so. We are with too few. In fact I have become truely
multilingual. I use R for most statistical stuff, microarrays and
(soon) deep sequencing. That is why I am driven to map to Ruby. To be
honest, these days I am also using BioPerl stuff. BioRuby has quite a
bit of functionality, but the BioPerl people are ahead and still
moving faster. We have a problem in bioinformatics. Too few people,
too many language domains.

> I mean, Python/Ruby/Perl seems that depends on the end user, they can do  
> the same things but for R ?

We can map the R C libraries using biolib. I have done the
fundamentals. Tell me what functions you require. I think mapping is the
way forward.

Pj.