expression_stat

[Instructions]

expression_stat
Find which genes are statistically different between groups defined by expression cutoffs or META value cutoffs.

If you divide samples into two groups, find the genes that are most different by t-test.

To use this method, first decide on cutoff values for samples less than a certain value for group A. Then set a cutoff for samples higher than a certain value for group B. Samples that do not meet either cutoff will be ignored. For each gene with expression data, do a t-test and record the t-statistic. Return the ranked list of genes sorted by abs(t-statistic).

group1 less than : lower expression cutoff (required) default 4
group2 greater than : higher expression cutff (required) default 10
num results : the number of results to return (default 25)
Plot heatmap : if checked, plot a heatmap of the returned genes.

HUGO: enter the gene symbol eg. EGFR. If using meta value cutoff use the format META-NNN to specify the meta value to create groups. Make sure the meta is numeric.
EXPRESSION : expression source (required)
META : set meta source and value to color samples on heatmap. (optional) Note: in order for the metavalue dropdown menu to show up, you must select the metasource and submit once so the menu can be populated from the source selection.

example Query. <9.5 >9.5, EXPRESSION = TCGA-BRCA-RSEM, META = TCGA-BRCA, VALUE = 3899 PAM50.RNASEQ, HUGO = RB1