From 71c0e2ad59d4ad0d2d951612c66e686c522b6979 Mon Sep 17 00:00:00 2001
From: hosseinvk <h.valipour.k@gmail.com>
Date: Mon, 10 Aug 2026 10:51:54 +1000
Subject: [PATCH] Correct GSEA mislabeling in g:Profiler (web) chapter

g:Profiler's "Ordered query" mode runs incremental ORA (hypergeometric
test) over prefixes of a ranked list; it does not implement the
Subramanian et al. 2005 GSEA algorithm (no running-sum statistic, no
weighting by ranking-metric magnitude, no permutation testing). This
was previously labelled "GSEA-style analysis" and "GSEA with
gProfiler", which risks confusing readers now that this chapter sits
right after the real GSEA(web) chapter. Retitled the section and
challenge, added a clarifying note contrasting the two methods, and
corrected the intro line's claim that gProfiler "supports both ORA
and GSEA".
---
 08-1-gprofiler-web.Rmd | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/08-1-gprofiler-web.Rmd b/08-1-gprofiler-web.Rmd
index 7a5d38c..7494f1a 100644
--- a/08-1-gprofiler-web.Rmd
+++ b/08-1-gprofiler-web.Rmd
@@ -7,7 +7,7 @@ Throughout the following chapters, each web-based tool is paired with its comman
 <!-- ## FEA in [g:Profiler](https://biit.cs.ut.ee/gprofiler/gost) -->
 ## <a href="https://biit.cs.ut.ee/gprofiler/gost" target="_blank"><img src="images/gp2-logo-new.png" alt="g:Profiler Logo" style="height:35px; vertical-align:middle;"></a>
 
-[gProfiler](https://biit.cs.ut.ee/gprofiler/gost) is known for its integration of numerous species and databases. It supports both ORA and GSEA, enabling users to assess Gene Ontology (GO), biological pathways, regulatory motifs and protein databases. With gProfiler one can
+[gProfiler](https://biit.cs.ut.ee/gprofiler/gost) is known for its integration of numerous species and databases. It supports ORA, including an ordered/ranked-list mode, enabling users to assess Gene Ontology (GO), biological pathways, regulatory motifs and protein databases. With gProfiler one can
 
 ### Steps to perform ORA in g:Profiler:
 
@@ -67,7 +67,10 @@ How is multi-query support implemented in gProfiler?
 
 How can one perform Under Representation Analysis in gProfiler?
 
-### Steps to perform GSEA-style analysis in g:Profiler:
+### Steps to use a ranked gene list in g:Profiler:
+
+g:Profiler's "Ordered query" option lets you analyse a ranked gene list, but it does **not** implement the Subramanian et al. 2005 GSEA algorithm covered in the GSEA (web) chapter earlier in the book. Instead, it runs standard ORA (hypergeometric/Fisher's exact test) on a series of prefixes of your ranked list (top 10, top 20, top 50 genes, and so on) and reports whichever prefix gives the strongest enrichment for each term. It does not use a running-sum enrichment score, does not weight genes by the magnitude of their ranking metric, and does not use permutation testing — so treat it as ORA over an incrementally-growing ranked list, not as an alternative implementation of GSEA.
+
 <span style="color:orange;">- Prepare Your Pre-ranked List:</span> Steps to provide a ranked gene list are given [here](degust.html).
 
 <span style="color:orange;">- Input Gene List:</span> Paste your prepared gene list directly into the input box on the g:Profiler web page or upload a file containing your list.
@@ -82,7 +85,7 @@ How can one perform Under Representation Analysis in gProfiler?
 
 <span style="color:orange;">- Run Query:</span> Same as above.
 
-#### **Challenge:** GSEA with gProfiler {- .challenge}
+#### **Challenge:** Ranked-list ORA with gProfiler {- .challenge}
 Download the Hallmark gene sets ([h.all.v2024.1.Hs.symbols.gmt](https://www.gsea-msigdb.org/)) from MSigDB and use it as Custom GMT.
 
 <!-- Full path to h.all.v2024.1.Hs.symbols.gmt -->