[R] precision of double in R

[Duncan Murdoch]

On 11/11/2005 9:55 AM, Burkhard Heil wrote:
> Hello Everybody!
> 
> 
> 
> I'm working on biological sequences and their alignements. I can't get
> around calculating something in the size of choose(1000,500) or 0.02^250.
> The result is either INF or 0 in the latter case. Is their any data type in
> R to calculate these things. Or some other solution?

Use logs.  log(0.02^250) is easy to calculate as 250*log(0.02).  There's 
also an lchoose() function; it gives

 > lchoose( 1000, 500)
[1] 689.4673

without complaining.

Duncan Murdoch