Ometer
Software for Multivariate Data Analysis for Genomics and Systems Biology
Summary
Ometer is software for multivariate analysis of functional genomics and systems biology data (it can be used for other types of data, too). Ometer is driven by a plain command line interface, which allows it to be called directly from scripts. Ometer is focused on efficient calculation of numerical results, which are written to tab-delimited files. Files are produced for visualization with gnuplot and graphviz, as well as reports formated in HTML.
Algorithms
- correlation analysis
- partial correlation analysis
- principal component analysis (PCA), including biplot display
- linear multiple discriminant analysis (MDA),
- linear multiple discriminant analysis with genetic algorithm variable selection - 2 different algorithms (GA-DFA and GA-DFA2).
- non-negative matrix factorization (NMF)
- factor analysis with varimax rotation (FA)
Download
ometer-0.60.tar.gz (source code 300.22 Kb)
To install, first download this file and untar it. Then run ./configure && make && make install, this will put ometer in /usr/local/bin. If you need to install ometer in other location, do instead ./configure --prefix=yourpath && make && make install.
version 0.60 for windows (1.92 Mb)
To install the windows version: download the executable file an save it in any folder that is in your path (eg C:\WINDOWS). If you are not able to do so, then put it in a new folder and change your path to include that folder (right-click on My Computer, select Properties, Advanced, and click on Environment Variables, then add it to the PATH variable) .
version 0.60 executable binary for Linux on Pentium4 (4.23 Mb)
To install only on a Intel 32 bit processor. Put this executable file on your path and change its the permissions to executable (chmod a+x ometer). Note that this will not run on other processors (eg on AMD). If you need to install and run on other processors, you should build ometer from source code.
Documentation
- File formats
- File transformations
- Correlation analysis
- Principal Components Analysis - PCA
- Linear Multiple Discriminant Analysis - MDA
- Linear Multiple Discriminant Analysis with GA Variable Selection - GA-DFA
Other documentation to appear later:
- Non-negative matrix factorization
- Factor analysis
- partial correlation analysis
- Producing plots of ometer results with gnuplot
- Producing correlation graphs with graphviz
Credits
Ometer was written by Pedro Mendes. The partial correlation algorithm was developed by Alberto de la Fuente. Other credits:
- CLO++, a command line option parser generator.
- LAPACK and BLAS routines for matrix algebra
- some binary versions include the Intel MKL version of LAPACK
- DCDFLIB, a library of comulative distribution functions, by Barry Brown, James Lovato, and Kathy Russell
- The C++ Standard Template Library, for data structures and algorithms
- MDA code based on original by F. Murtagh <murtagh@scivax.stsci.edu>, Munich, Jan 1986
