TagSNPFinder USER Guider
Version 1.0
Zhenqiu Liu and Shili Lin

Introduction

Multilocus LD measure and Tag SNP selection are very important in human genetics study.  TagSNPFinder defined a novel multilocus LD measure ER with generalized mutual information. We also designed  a general scheme to select 'tagging' SNPs through  combining the association and diversity together. Currently TagSNPFinder have routines to select Tagging SNPs with exhaustive search forward, backward, and stepwise selection.  The Matlab verision of both ER and  are both implemented.

The routines for identifying and evaluation tagging SNPs can be applied to haplotype data from any source (that is, haplotypes determined by phase inference in singleton samples or determined in family structured samples).  


The guide is written on the assumption that MATLAB has been installed either on  linux or PC.  Note that a free MATLAB look-a-like package called Octave is also available. 
Web-page for MATLAB is:
http://www.mathworks.com/products/matlab/

Web-page for Octave is:
 http://www.octave.org/


Set up

Select a directory and put the .p or .m file in that directory and run the Matlab.  We did not set subdirectories for simplity. 

Functions
 backsel.m:  perform the backward selection of SNPs with conditional    entropy.
bwdsel.m: select the SNPs backward with ER and H(S) together.
fwdsel.m: select the SNPs forward with ER and H(S) together.
stepsel.m: selct tagging SNPs with forward and backward selection  together.

 myperforms.m: perform exhaustive search for SNP selection.
  
multinf.m: Multilocus LD measure of ER.
Mutinf.m:  Multilocus LD measure of ER.

Mylds.m:     Matlab version of the multilocus LD measure epison.

Other utility functions:
entropy.m: measure the entropy of single locus.
Jointinf.m: measure the joint entropy of a SNP set.
Parsinp.m: parsing the input data into right format.

Test Examples
medriv.m  To show how  ER works. To running  it, just type medriv at the command line of Matlab.
fsdriv.m: To show how  forward selection works for Tagging SNPs. To
run it simply type fsdriv at the command line. 
exhdriv.m: Example for exhaustive search. 

Datasets:

Clayton.txt 
examples ...

Reference:
[1]. Zhenqiu Liu, Shili Lin:Multilocus LD Measure and SNP Selection with 
Generalized Mutual Information 1/07/2005. Submitted.
[2] TagIT. Version 1.17.  2003. Mike Weale and David Goldstein. http://popgen.biol.ucl.ac.uk/software.html


Updating:
Currently only P codes are available at our website. The  source codes  (.m)  are provided upon request. If you are interested,  please send us a email at
shili@stat.ohio-state.edu
or 
liu@stat.ohio-state.edu

C++ version will be available soon....

 
