Welcome to qubic’s documentation!

The qubic download page can be accessed at Project Homepage.

Usage

This software provides a biclustering module for microarray data. For a set of genes and a set of conditions, the program outputs a block-like structure which shows uniform pattern within the block, the block would contain only subsets of all given genes under subsets of all given conditions.

Certain parts of the code uses open-source data structure library codes, including:

• fib, copyright information in fib.c
• Mark A. Weiss’s data structure codes

Installation

Note

qubic currently will only run on linux platform, as it is dependent on GNU function.

Simply put “qubic0.1.tar.gz” in any directory:

$ tar zxvf qubic0.1.tar.gz

enter the folder qubic0.1 and type make, the compiled codes are within the same directory as the source.

Note that in folder 'qubic0.1', there is a folder 'bin' which contains the program complied on a 64-bits machine, and all of the results in our paper are tested on a 64-bit machine.

Inputs and outputs

The major program in the provided package is qubic, it can parse two formats of files, discrete data and continuous data, and examples for each are provided. See help and look at all available options.

$ ./qubic -h

Take a look at toy_example (discrete data) first. And try to run clustering:

$ ./qubic -i toy_example -d

-d is important here since it tells the program that this is discrete data.

Then look at a larger file yeast_example (continuous data). Try to run:

$ ./qubic -i Ecoli -f .25

This restricts no two blocks overlap more than 0.25 of the size of each one. And the other parameters are default value (see details by typing './qubic -h').

For each input file, our program generates three output files, namely,.blocks file, .chars file and .rules file. In .blocks file, you can see all the biclusters the program found, especially, we use a blank line to separate the positively and the negatively (if any) correlated genes in each bicluster. As to .chars file, it provides the qualitative matrix of the microarray data to usrs with some details of how to discrete the data in .rules file. You can find further details about how to represent a microarray dataset with a qualitative matrix in our paper.

Changelog

• Version 0.1 prototype of algorithm
• Version 0.2
  new function that generate the biclusters (if any) contains the genes usrs provided: usrs can put the genes at hand in a 'genes' file (see details of format in 'genes' file in current folder), and type 'perl expand', then you will get a '.expansion ' file for each group of genes in 'genes' file.
  new function that can expand the current blocks generated by our program in specific environment: First, type './qubic -i Ecoli', then type './qubic -i CRP', finally type './qubic -i Ecoli.chars -b CRP.blocks -s' and the program will generate a 'CRP.blocks.e xpansion' file containning the enlarged blocks in CRP.blocks. You also can type "make test" to realize the above three commands.

Contact

Any questions, problems, bugs are welcome and should be dumped to
Haibao Tang <bao@uga.edu>

Qin Ma <maqin@csbl.bmb.uga.edu>