Impute Missing Values. Supplement to:

LSimpute: Accurate estimation of missing values in microarray data
with least squares methods. Trond Hellem Bø, Bjarte Dysvik and Inge
Jonassen. Nucleic Acids Research 32 (3): e34

Corresponding author: trondb@ii.uib.no

USAGE:

To run the program, use the following command line arguments:

"java -jar -server LSimpute.jar <InputFile> <OutputFile> <Method>"

<InputFile> The source with missing values where missing values is any non-number string (like "NULL"). 
Strings you can NOT use for missing values are the strings "NaN" and "infinity".
<OutputFile> The name of the result file
<Method> Defines the method to use for imputation, one of the following: 
<Method> 0 : Row mean 
<Method> 1 : EMimpute_gene
<Method> 2 : EMimpute_array
<Method> 3 : LSimpute_gene
<Method> 4 : LSimpute_array
<Method> 5 : LSimpute_combined
<Method> 6 : LSimpute_adaptive

Please refer to the article for explanation of the different
imputation methods.



INPUT FILE FORMAT:

The intput file must be in the following format:
First row is column identifiers with a uninportant first element,
separated by tabs. Subsequent rows contain a row identifier as first
element and data/missing values in the other, separated by tabs.

[  --  ]	[ColID1]	[ColID2]	[ColID3]	[...]
[RowID1]	[ Data ]	[ Data ]	[ Data ]	[...]
[RowID2]	[ Data ]	[ Data ]	[ Data ]	[...]
[RowID3]	[ Data ]	[ Data ]	[ Data ]	[...]
[  ... ]	[  ... ]	[  ... ]	[  ... ]	[...]