statistics of DNA miroarrays

Does anybody has a clue?

Spotted cDNA arrays–not oligos.
I’m famaliar with Affymetrix’s software.

What’s a good package for cDNA arrays?

Is there a package people who have dealt with, like–?

I realize my question should be-who has done stats on cDNA arrays.
And what are those stats?

What do you mean by stats? Hits? False positives? Total numbers?

Use R and the packages at Bioconductor. The vignettes are very helpful. Also have a look at Terry Speed’s page.

Sorry, jk–I should have been more explicit, although I’m not sure exactly what you mean by hits with microarrays (in a way, these are user-defined --and that’s what I’m asking about). False positives are a danger with any statistical analysis. I’m looking at more Northerns and RT-PCRs than I care to think about.

I get a .tif file back from our Core facility and I use TIGR’s Spotfire to turn intensities into quantifiable numbers. I’m comparing 2 levels of treatment and want to analyze the ratios.

So, now I have a spreadsheet with thousands of numbers. Now with ratios, I get that I should Log transform the data–under-expressed genes will have a ratio of less than 1 and over-expressed genes will range from 1 to (theoretically) infinity.

What I wanted to know were things like: how do you determine background (using control spots), what manipulations do you use to correct for this, at what (corrected) ratio do you consider a gene differentially expressed, and what confirms these genes statistically?

What are the current ‘standards’ and what software makes this easy?

I realize this is a fairly new field and different papers do all this differently. The first microarray papers got published just describing genes that were up or down regulated, say ‘3 or 2-fold’, using raw ratios.

Bleach—thanks for the links. I’m actually familiar with some of Terry Speed’s work, although I didn’t have that particular webpage.