[BioC] Alert to developers about changes to edgeR

[Gordon K Smyth]

The deadline for the 2.10 Bioconductor release is very close.  There are a 
few packages that depend on or suggest edgeR.  This email is to alert 
developers to changes in edgeR that might potentially break code.  One of 
the changes below was made only two weeks ago, so I thought it would be 
worth an email to alert people.

1. estimateCommonDisp() and estimateGLMCommonDisp() now return 
dispersion=NA when there is no replication.  This is to force users to 
make a decision for themselves about what action to take in this case. 
This means that any subsequent function that uses the common dispersion 
will fail when there is no replication.  If you want to reproduce earlier 
behaviour, you need to explicitly enter dispersion=0 to subsequent 
functions.

2. The column headings in the table returned by exactTest() and glmLRT() 
have changed.  For example, the column called logConc is now logCPM (log 
counts-per-million).  If anyone is pulling out columns such as logConc by 
name, such code might be broken.

There are lots of other changes since the last release, but the above two 
are fairly recent and the most likely to break dependent code.

Best wishes
Gordon

---------------------------------------------
Professor Gordon K Smyth,
Bioinformatics Division,
Walter and Eliza Hall Institute of Medical Research,
1G Royal Parade, Parkville, Vic 3052, Australia.
smyth at wehi.edu.au
http://www.wehi.edu.au
http://www.statsci.org/smyth

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}