Version 0.8.6

[Davide Cittaro]

Hi there,
I have really few time to spend on dspchip improvements, nevertheless I've added some features in version 0.8.6.
- modified --dist option: now it's boolean and tries to infer the best distribution model among normal, lognormal, powerlaw and gamma (most likely will be the last one... I personally believe gamma distribution may be the proper model for enrichments... but this is another story)
- modified --scale option: instead of scaling data in specified intervals, dspchip tries to restore an original scale. Too many times I had people complaining about low values after denoising (what does 0.00231 mean after all?).
- some bugfixes: I've noticed bedgraph and bigwig contained duplicated values... this is probably an old bug that has been harmless until new bigWig converters started complaining :-(

Some nice facts:
dspchip user group has really few members, dspchip has been presented in poster sessions at BITS 2011 (for Italy) and HitSeq 2011. This said, there have been more downloads than expected... Probably somebody spread the word, I think most of the people land on dspchip pages looking for dsp processors (engineers...) though :-)
d

--
Davide Cittaro
daweo...@gmail.com
http://daweonline.googlepages.com/