[BioC] Nimblegen chip data analysis

Todd Richmond todd at nimblegen.com
Mon Aug 30 17:19:51 CEST 2004


Sean Davis wrote:
> Andrew,
> 
> Nimblegen arrays are simply two-color arrays on (legal) steroids.  You can
> simply treat them as such.  

Sorry, I feel like I should jump in and correct this. Most NimbleGen 
arrays are one-color arrays, similar to Affymetrix arrays. We do some 
two-color arrays, but most of it is not for expression studies.

> One thing to be careful of, though, is what you assume about
> differential expression if you have a highly customized array. There
> aren't methods to read Nimblegen data, so you will probably have to
> read it with standard R functions (read.table, for example). Decide
> what hypotheses make sense for your data and think about what makes
> sense for normalization (again, depends on the chip design, to some
> degree).

Due to the high level of customization for our array designs, we haven't 
invested the time to develop a specialized set of tools to load 
NimbleGen data into Bioconductor. I read in data with read.table and 
pull out row/columns as needed. Most of the Bioconductor packages have 
lower level functions that allow you to pass a matrix or dataframe as an 
argument, instead of the typical AffyBatch or expression set. It's a 
little more work, and sometimes a bit painful, but it does allow for 
lots of flexibility. Sean's point about normalization is a good one - 
depending on the application, you'll find that some methods of 
normalization are more appropriate than others. It might take a bit of 
experimentation to figure out which normalization method is best suited 
to your design/samples/application.

Regards, Todd


> 
> Sean
> 
> On 8/27/04 17:51, "Andrew Kern" <adkern at ucdavis.edu> wrote:
> 
> 
>>has anyone out there used bioconductor to analyze nimblegen
>>oligonucleotide chip data? if so could you point me in the right
>>direction to get me started? i've never before analyzed microarray data
>>and i'm hoping to use bioconductor for this project.
>>cheers,
>>andrew kern
>>

-- 
******************************************
Todd Richmond, PhD
Manager of Research Informatics
NimbleGen Systems, Inc.
Email: todd at nimblegen.com
Phone: 1-608-218-7651


More information about the Bioconductor mailing list