I'm reading Subramanian et. al's (2005) original GSEA paper. Because GSEA increases the running sum statistic the most when genes appear close to each other (and hence yields a large ES score), it seems to me that it assumes that pathways (as defined by prior gene lists) involve genes that are positively correlated with each other. For example, if we had a moderate number of genes at the top and bottom of the list, then we would NOT detect a pathway that involves genes that are both at the top and bottom (i.e genes that are negatively correlated with each other). Am I understanding this correctly?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.