[Bioc-devel] Affy QA function for large chip set

Hi Ben,

MAplot with plot.method = "smoothScatter" saves a small footprint in PDF file, and rendering of plot is fast. I also likes the type argument which controls the plot of pm, mm or both.

Another question,

In the AffyPLM package, many QA graphs are bi-products of RMA model fitting. If I like to see the same type of QA graphs for gcRMA model fitting, is there any way to go?

Thanks,

Ping




-----Original Message-----
From: Ben Bolstad [mailto:bmb at bmbolstad.com]
Sent: Friday, June 22, 2007 6:14 AM
To: Ling, Ping [ALZUS]
Cc: bioc-devel at stat.math.ethz.ch
Subject: Re: [Bioc-devel] Affy QA function for large chip set

#3. ma.plot function in the affy package is pretty slow with probe level data, it renders a lot of data points in the output. When we come to export 100 MA plots with 6 MA plots per page in a PDF file, the file gets really big and navigation between pages becomes difficult.   I used a local weighted data point reduction sampling method to keep the contour and density profile of the MA plot while significantly reduce the data points. (I used such a method to retain similiar visual perspective with 5% data points).

you might want to consider using the "smoothScatter" plotting method,
rather than just letting it plot all points with probe level data. This
and other potential options for MAplot (which typically wraps ma.plot)
are described here (though were it says exprSet, it should really say
ExpressionSet):

http://bioconductor.org/packages/bioc/vignettes/affyPLM/inst/doc/MAplots.pdf