Question

Warning messages: gseaKEGG

0

Entering edit mode

adR ▴ 40

@do-it-23093

Last seen 5 months ago

Germany, München

Hello Dear friends,

I had a question and was already posted in the Biostars platform but recommended by a friend to bring it here for further feedback. The questions are here bellow in the link. https://www.biostars.org/p/454938/#454949 Could you please give me some more tips on how to handle these? The input data has only geneIDs(enterZ) and log2fold change values of the genes. Thank you so much!

clusterProfiler r Tutorial • 1.5k views

ADD COMMENT • link updated 4.4 years ago by Kevin Blighe ★ 4.0k • written 4.4 years ago by adR ▴ 40

score 0 · Answer 1 · 2020-08-15

0

Entering edit mode

Kevin Blighe ★ 4.0k

@kevin

Last seen 7 days ago

Republic of Ireland

The warning messages say more about your upstream analysis, as opposed to this being an issue with clusterProfiler. You may want to figure out why ties may exist in your [assumed ranked] input data. Ties of this nature may exist in studies with small sample n, imbalanced comparison groups, or, generally studies of poor design and/or poor model assumptions.

If you want to show a reproducible example, that would help us very much.

Kevin

ADD COMMENT • link 4.4 years ago Kevin Blighe ★ 4.0k

0

Entering edit mode

Thanks, Kevin, I would like to share with you the top 50 in the following picture. In the picture, you see the named gene with the log2FC values. The code and link for the picture are below I used:

z <- gseKEGG(geneList = y, 
    pvalueCutoff = 0.5, # pvalue cutoff 
    organism = "hsa", # supported organisms listed below
    keyType = "kegg",
    exponent = 1,
    eps = 0,
    pAdjustMethod = "none",
    minGSSize = 0,
    by = "fgsea",
    use_internal_data = FALSE)

top 50 genes with their LFC values

ADD REPLY • link 4.4 years ago adR ▴ 40