Figure 4

Simple examples for the term significance test. Each table represents the (hypothetic) data for one test term. The second column shows the count of the test term in the document set of a gene, and the third column shows the expected count according to the null distribution (assuming that the term is not related to the gene). The expected count is the product of the frequency of the term in the background collection and the length of the document set of the gene. E.g. in the first row of table (A), 5 means the term appears five times in all the documents associated to g1, and 0.1 is the expected counts according to the background. (A) An example where the term may be related to the first two genes. (B) An example where the term does not appear to be significantly related to any gene.