Significance Analysis of Microarrays implementation using Matlab
Microarray experiments enable the simultaneous measure of expression levels of large amount of genes and have many applications. A widespread one is finding set of genes that are differential expressed. Significance Analysis of Microarrays (SAM) helps to produce those sets using multiple testing techniques. There is unfortunately not yet a public tool enabling to do SAM using the Matlab platform. We here define MatSAM, a SAM implementation in Matlab, and show that it yields results of high confidence comparatively to those obtained by putative tools available in the R programming environment. MatSAM can be used in conjunction with Matlab Bioinformatics toolbox to perform further analysis. MatSAM is based on the BioConductor project package siggenes (Schwender, 2012), courtesy to siggenes maintainer' permission.
Contents
Disclaimer
This is the documentation home page. To browse the API documentation, one should use the links provided below.
This is an algorithm implementation documentation and not a software's, that is the documentation here provided does not deal with using Matlab or any of its toolboxes. Once one has installed Matlab along with the Statistics and Bioinformatics toolboxes, they could use MatSAM in the same way as for any set of scripts ran in Matlab.
Download
MatSAM source code and supplementary files can be found here.
Demonstration
To see a full demonstration of the standard way MatSAM is ran, please refer to the demonstration page.
Manual pages
GLOBAL_ARGS declares and defines global arguments.
SAM_INPUT reads the raw files containing gene intensities and gene names.
SAM_FUNC proceeds to the Significance Analysis of Microarrays.
D_STAT_FUNC corrects the D-scores by the fudge factor.
PURGE_NAN removes all NaNs in the data.
TEST_TYPE decides which statistical test to launch.
GET_FOLD_CHANGES computes the observations fold changes.
STATEST_NUM_DENUM gives the numerator and the denominator of a statistical test.
FUDGE_FACTOR determines the fudge factor.
SETUP_MAT_SAMP randomly permutes the experiments.
D_NULL computes D-bar scores and P-values after statistical testing the data.
BUILD_DPERM computes D-scores for a given class permutation.
BAL_PERM divides an experiment permutation following provided classes design.
PERM_STATEST_NUM_DENUM computes the numerator and the denominator of the required statistical test.
PI0_ESTIMATION estimates the prior that an observation is expressed.
QVALUE_CAL computes the q-value for each observation.
RANKING ascendantly ranks the data.
FULL_RESULTS prints out summary of the SAM results.
STATS_CAL gathers the statistics of the SAM.
SAM_RESULTS gathers all the SAM results.
SAM_PLOT yields the SAM Q-Q plot.
SAVE_SAM_RESULTS writes the SAM results in a file.
License
MatSAM is inspired by the siggenes package (Schwender, 2012). siggenes is a free software distributed under the LGPL version 2 or later (GNU Lesser General Public License).
MatSAM is distributed under the BSD License, no advertising (Berkeley Software Distribution).
References
- Virginia Goss Tusher, Robert Tibshirani, Gilbert Chu (2001). Significance analysis of microarrays applied to the ionizing radiation response. Proceedings of the National Academy of Sciences 98(9): 5116-5121.
- Holger Schwender (2012). siggenes : Multiple testing using SAM and Efron's empirical Bayes approaches. R package, version 1.38.0.