[GeneX-dev] Re: CyberT questions...

[Harry M. <man...@ho...>]

> Michael Pear wrote:
> 
> Hi Harry,
>  I've massaged the data here to analyze the dataset with CyberT. I'm left with a few questions
> that I hope you can help with.
> 
> 1) I'm left with the impression that there was a problem with the CyberT scripts that you are addressing
> to ensure negatives cannot get passed to cyberT. Is that correct? That would certainly make operation
> more bulletproof.

Yes, I'm patching code now that detects negs and scales the data into the + range for use with CyberT.

> 2) There appears to be a "minrep" level set to 2 when calling "doitall" that is to prevent t-testing on too
> small a number of replicates. However, it seems that there is still a "p" value and fold computed for  spots
> that violate this minimum requirement...am I missing something? We are finding the some of our best
> "p" values are for spots that do not satisfy the "minrep" requirement.

Hmm - that shouldn't be the case - it should refuse to calculate these.  No wonder that they're the best p's are with these - they'd tend to have less variability.  This is with the C+E cybert?

> 3) Is there any description of the details of the Bayesian approach or other formulas used in cyberT?
> ---other than the R code, which I haven't had a chance to really take on and understand....not ANOTHER
> language, pleeeease....I'm reaching saturation :-(

Yes, I'm attaching a paper that described it nasty detail.  It's not that bad actually.  I'm in sympathy with your take on R - it's nice, but it is ANOTHER language - that's why I'm doing as much of
the data management in perl as possible :)

> 
> 4) I'm wondering about including the full description for genes in the cyberT output and xgobi (for identification).
> I would expect it to be a rather simple change to have the output data include the description rather than
> gene name...any reason to expect that this mod would give problems? Do you have any plans in this regard
> for future versions?

Not with xgobi (without recoding a bit - to include both label and description), but with ggobi I've been assured that this will be possible.  This *IS* one of the key things that I'd like to address
in the short term, one way or another.  If it takes recoding xgobi, I'll do it, but ggobi claims that it'll be possible.
http://www.ggobi.org

This may also be possible with J-Express - just posted a questio to their list about this..




> 
> Regards,
> 
> Michael Pear
> 

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...