This is a readme file for using makeCGI package to annotate
CpG island given sequence. 

I. Input:
The makeCGI package takes two types of input for sequence: 
text file or BSgenome package. The type is specified in
the configuration file by setting .CGIoptions$rawdat.type
to be "txt" (for text file) or "BSgenome" (for BSgenome package).

II. Steps to annotate CGI:
0. Specify the parameters for CGI finding. For details 
please look at the sample configuration file. 

1. Call "count.CG" function to count the numbers of C,G
and CG in each L bp window.

2. call "calc.postprob" to calculate the posterior probabilities
of being in CGI for each L bp window. Result from the function
is a data frame of four columns for chromosome, position, 
posterior probabilities for CpG chain and posterior probabilities
for GC content chain. The result will be saved as rda files 
(one per chromosome) to designated locations.

3. With the posterior probabilities and a threshold for 
CGI, call "makeCGI" function to annotate CGI. 

Note that steps 1, 2 and 3 are sequential and can run separately.
For example, to annotate CGI for another threshold, one can 
rerun step 3 without rerunning  steps 1 and 2.

This program requires a lot of memory, depending on
the size of the genome being analyzed. The computational time
could be substantial. Computer clusters should be used if
available. A parallizable version of the package is under
development.

There are two examples available for C. elegans (sequence in BSgenome)
and E. coli (sequence saved in text file). Go to the test 
directories and run "script.R" to annotate CGI for them.