Wigwams in MATLAB Load one of the datasets ---------------------------------- >> load ../datasets/ArabidopsisAbioticShoot other datasets: ArabidopsisAbioticRoot EcoliDiauxie HomoSapiensMS YeastCellCycle YeastStress run Wigwams ------------------ MultiHyperAll(dataset, geneList, pseudo_set_sizes, job, score_threshold) -dataset = dataset you have loaded -geneList = list gene indices you wish to run Wigwams on, where Wigwams will iterative test each gene seperately. - list_set_sizes = default is [50:50:250]. Defined as L in the paper. -job = a name to call your output e.g. '_job1' -score_threshold = if you are using the differential expression scores (not in datasets mentioned above), pick a cutoff, above which, genes are considered DE. If not, type 'no'. >>MultiHyperAll(ArabidopsisAbioticShoot, 1:6, [50:50:250], '_genes-1-6', 'no') run Pruning procedure ------------------------------ pruning2(dataset, pvalue_cutoff, correlation_cutoff, significant_difference) Dataset as before. pvalue_cutoff = log p-values below this value considered correlation_cutoff = cutoff below which pairs of cluster are deemed unsimilar significant_difference = difference by which one log p-value is smaller than another log p-value to be considered significantly stronger. See Figure 2 for explanantion. >>pruning2(ArabidopsisAbioticShoot, -10, 0.75, 25) export the overlaps -------------------------- overlap_function(dataset, pseudo_set_sizes, final_genes, score_threshold) dataset, pseudo_set_sizes and score_threshold as before. final_genes = the gene indices from the pruning stage. In order to find these indices, open 'final_genes.txt', which will be located in the 'output' folder. In the second column, each cell will say something like 'gene number (e.g. 6) pruned 75 genes'. The gene number (e.g. 6) has been kept, along with the cluster it has seeded. Therefore, to get the final gene indices, copy and paste all the gene indices in that column into a vector: >>final_genes = [copy and paste here]; >>overlap_function(ArabidopsisAbioticShoot, [50:50:250], 6, 'no') overlap_function will create three files: output.eps = all plots of significant clusters overlap_table.txt = all clusters, datasets they are significantly co-opted in, number of datasets co-opted in, and size of cluster. page_two.txt = each cluster and gene members of cluster.