ctlab / fgsea

Fast Gene Set Enrichment Analysis
Other
379 stars 67 forks source link

scoreType question #79

Closed pstew closed 4 years ago

pstew commented 4 years ago

How are the ranks treated for scoreType = "pos" ? At one point in the code you sort by decreasing values. For scoreType = "pos", does this mean the largest value is the first rank? I just want to make sure the smallest value is not the first rank (e.g. it is not counting upwards from 1 for the ranks). Does this then mean the smallest value is the first rank for scoreType = "neg"? Thanks!

assaron commented 4 years ago

Ranks are treated in the same way, independent of scoreType: they are always sorted from largest positive to largest negative values. What scoreType does is changes which extremum of the enrichment curve is used as an enrichment score value: pos -- for maximum (always non-negative), neg -- for minimum (always non-positive), std -- whichever from max and min has the greatest absolute value. Consequently, the p-value calculation is changed a bit, with pos and neg leading to single-end p-value, as opposed to the sort-of two-tailed standard P-value.

pstew commented 4 years ago

Thank you! I really appreciate the quick response. Thank you for the great, easy to use package!