NEWS

LPE 1.6.0

Changed the name of the function "mt.rawp2adjp" to "mt.rawp2adjp.LPE" in order to avoid conflict with the original function from multtest library.
Added the references to paired LPE - "LPEP library" and LPE for multiple conditions - "HEM library".
Added an option "Bonferroni" in fdr.adjust.R to get Bonferroni adjusted p-values. (Though it is recommended to use FDR, Bonferroni adjusted method has been added here for users who want it.)
Updated the document 'LPE.pdf' (hands on demonstration of LPE) in inst/doc. Changes summary:

Created a project called "r-lpe" on sourceforge to get the most recent files. One can keep track of changes and checkout the latest version by anonymously checking out lpe from sf.net: cvs -z3 -d:pserver:[email protected]:/cvsroot/r-lpe co -P LPE
Bug fix in baseOlig.error.step1 and baseOlig.error.step2: For some data sets, adjacent quantile values were same due to thresholding/nature of the data - which caused the number of genes to be selected for var(M) caluclation as 0, and hence there was an error: "Error in var(x) : 'x' is empty".
Added a check in baseOlig.error.step1 and baseOlig.error.step2 to see if min. value of variance does NOT go negative.
Updated the email address. New address is: <[email protected]>
Changed the default value of the parameter probe.set.name to NULL, so that if the GeneIDs are not provided, then rownames (1,2,3,...) are considered as GeneIDs.
Changed the default argument of preprocess function 'lowess=TRUE' to FALSE. User should specify if they need lowess transformation.
Added a reference (new published paper on "rank-invariant resampling for FDR calculations.

More robust detection of outliers in two sample comparison. (Thanks to HyungJun Cho for the correction).
Function lpe.R is broken in several small functions, easier to understand.
Built under R 2.0.0