Method to find pathways different between 2 groups
2
0
Entering edit mode
Chris ▴ 10
@3fdb6f97
Last seen 3 days ago
United States

Hi all,

I have two groups would like to compare which pathways different between them. I ran gene set variation analysis follow up with limma, fgsea, camera. enrichGO() also rank the pathway but we can't choose which gene set we want to analyze. Each of them gives different result. Haven't tried Qiugen IPA yet. Could we combine the result from each tool? Would you please suggest a method in this case? As Gordon comment on this post Why does GSEA on edgeR results for randomized samples give highly significant p-values?, seem p valule from fgsea is not truly significant as it is. When I search camera limma on pubmed, I found 2 papers. Thank you so much!

fgsea GSVA CAMERA • 375 views
ADD COMMENT
2
Entering edit mode
Robert Castelo ★ 3.3k
@rcastelo
Last seen 2 hours ago
Barcelona/Universitat Pompeu Fabra

Hi, regarding how could you combine the result from each tool, one option is using a so-called "ensembl method", for instance, the package EGSEA, described in depth in the publication by Alhamdoosh et al. (2017), which includes camera and GSVA, among others, as optional tools to include in the analysis. If you have questions about the EGSEA package, please remember to add the tag EGSEA to the question.

ADD COMMENT
1
Entering edit mode
@gordon-smyth
Last seen 1 hour ago
WEHI, Melbourne, Australia

When I search camera limma on pubmed, I found 2 papers

I am unclear what point you are trying to make. limma's camera method has been formally cited in nearly 800 published papers.

ADD COMMENT
0
Entering edit mode

Sorry for not clear! Glad to know camera() is widely used. I just try to find papers use this method for learning purpose. Which gives more context and detail examples, so I may know when to use this method. This paper compare some pathway analysis methods include camera: https://pubmed.ncbi.nlm.nih.gov/29881462/ The conclusion is sigPathway and GSEA are the most powerful methods. There is a post about similar question years ago which I think still haven't had a clear answer. Statistic for GSA

ADD REPLY
1
Entering edit mode

sigPathway is another method that does not account for inter-gene correlation and which gives inflated significance, as we showed in our 2012 paper on camera.

ADD REPLY
0
Entering edit mode

Thanks Gordon! Don't know why follow by email doesn't work which automatically switch to follow by message.

ADD REPLY

Login before adding your answer.

Traffic: 782 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6