Send gene lists to over-representation, not ReactomeGSA - #300
Merged
Merged
Conversation
The user-guide prompt said "gene set analysis is ReactomeGSA" with no
word about gene lists, overriding the guide itself, which files a pasted
identifier list under Analysis Data (over-representation) and ReactomeGSA
under Analysis Gene Expression. Reported 2026-09-29: "here is my gene
list TP53, ERBB3 and JAX9" was told to paste it into ReactomeGSA.
The answer sweep required that wrong answer: its gene-list question had
must=("ReactomeGSA",), so a correct answer would have failed a deploy.
- Prompt: one added line -- a list of genes is over-representation via
Analyze Data; never send a gene list to ReactomeGSA. The original
ReactomeGSA sentence is kept, URL included: measured, removing
reactome.org/gsa or restructuring the rule made GSEA answers mix in
ReactomeFIViz's Cytoscape steps in 6/6 runs (1/6 before). The URL is
real: beta.reactome.org/gsa redirects to the quantitative analysis tab.
- Sweep: the gene-list question now expects over-representation, and the
reported message is added.
Measured, 5 runs each: old prompt 0/10 on the corrected checks, new 10/10.
GSEA answers mixing FIViz steps: 1/6 before, 2/6 after (n=6, noise). Full
sweep 16/16.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Reported on 2026-09-29: given "here is my gene list TP53, ERBB3 and JAX9", the model said to paste the list into ReactomeGSA.
Cause
must=("ReactomeGSA",), so a correct answer would have failed a deploy.Change
Prompt: one added line. A gene list is over-representation, run through Analyze Data; never send a gene list to ReactomeGSA. The original ReactomeGSA sentence is kept, URL included, because the alternatives measured worse:
reactome.org/gsais real:beta.reactome.org/gsaredirects to the Pathway Browser's quantitative analysis tab.Sweep: the gene-list question now expects over-representation, and the reported message has been added as a new question.
Verified
./checks.shpasses.Note
The prompt is shared with the search page, which gets the same correction.
The GSEA answer mixing in FIViz steps happens on the old prompt too (1/6). It is not gated: at that rate a sweep check would fail at random, and a gate that fails at random teaches people to ignore it.
🤖 Generated with Claude Code