We have applied Support Vector Machines (SVM) and a stratified cross‐validation approach to analyze a large microarray experiment dataset from D. melanogaster in order to predict possible functions for previously un‐annotated genes.
We have applied Support Vector Machines (SVM), a sigmoid fitting function and a stratified cross‐validation approach to analyze a large microarray experiment dataset from Drosophila melanogaster in order to predict possible functions for previously un‐annotated genes.
Consequently, 80 human, mouse and rat microarray experiments (datasets) from Gene Expression Omnibus (GEO) database met the selection criteria.
Microarray experiments (dataset VIII) were performed with custom UNC Agilent two-channel microarrays.
Basing on the microarray experiment (the datasets that we chose from GEO were from good quality and already normalized) the fold change between the two conditions can be computed for each gene.
In our study, we were only able to obtain the negative interaction data from a microarray experiment although the dataset had an imbalanced problem.
All datasets in this study were published within the past 5 years (2004-2009) and following the Minimum Information about a Microarray Experiment (MIAME) guidelines, including 49 datasets, with 3,832 samples containing 16, 15, 11, and 7 datasets from colon, liver, lung and thyroid cancer, respectively (Additional file 1: Table S2).
Validation of new findings with independent data is essential for us to believe that the results are indeed 'real', and this has been encouraged and facilitated by making it a prerequisite for publication that gene expression datasets are publicly accessible according to the MIAME (Minimum Information About a Microarray Experiment) guidelines [ 28], allowing complete datasets to be downloaded.
Supplementary information on each of the microarray experiments and all datasets are available online [ 22, 23].
Along the attempts to search large public microarray databases at experiment dataset level, individual gene expression profile level search method has also been conceptualized and developed [ 8].
