hi,

Such generalized view of the online tool used would not help. Each have a database (or multiple) like Gene ontology classifications or similar customized annotations which are used with different types of statistical test. At the end each will come up with different results.

My experience is use any well documented tool and check what statistics they use. For large gene lists you must take pathways/ processes that pass multiple testing correction.

Choose one reasonably and stick with it to test if your hypothesis is making sense.

See my comment to your other post.