Problem description:
data.df <- bitr(data2, fromType = "SYMBOL", toType = c("ENTREZID", "ENSEMBL"), OrgDb = "org.Mm.eg.db")
select()' returned 1:many mapping between keys and columns
Warning message:
In bitr(data2, fromType = "SYMBOL", toType = c("ENTREZID", "ENSEMBL"), :
19.54% of input gene IDs are fail to map...
System information
R version 3.5.1 (2018-07-02)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.1 LTS
You'll get better help by including a reproducible example, called a reprex
The message select()' returned 1:many mapping between keys and columns is normal. In fact one of examples in the documentation uses syntax identical to yours
ids <- bitr(x, fromType="SYMBOL", toType=c("UNIPROT", "ENSEMBL"), OrgDb="org.Hs.eg.db")
(where both of the toType arguments are returned by keytypes(org.Hs.eg.db) as are yours
The second message arises from the contents of your data2 object. Confirm that you have created data.df with
head(data.df)
> data.df <- bitr(data2, fromType="SYMBOL", toType=c("ENTREZID", "ENSEMBL"), OrgDb="org.Hs.eg.db")
'select()' returned 1:many mapping between keys and columns
head(data.df)
SYMBOL ENTREZID ENSEMBL
1 GPX3 2878 ENSG00000211445
2 GLRX 2745 ENSG00000173221
3 LBP 3929 ENSG00000129988
4 CRYAB 1410 ENSG00000109846
5 DEFB1 1672 ENSG00000164825
6 DEFB1 1672 ENSG00000284881
I used the documentation for data2
data2
[1] "GPX3" "GLRX" "LBP" "CRYAB" "DEFB1" "HCLS1" "SOD2" "HSPA2" "ORM1"
[10] "IGFBP1" "PTHLH" "GPC3" "IGFBP3" "TOB1" "MITF" "NDRG1" "NR1H4" "FGFR3"
[19] "PVR" "IL6" "PTPRM" "ERBB2" "NID2" "LAMB1" "COMP" "PLS3" "MCAM"
[28] "SPP1" "LAMC1" "COL4A2" "COL4A1" "MYOC" "ANXA4" "TFPI2" "CST6" "SLPI"
[37] "TIMP2" "CPM" "GGT1" "NNMT" "MAL" "EEF1A2" "HGD" "TCN2" "CDA"
[46] "PCCA" "CRYM" "PDXK" "STC1" "WARS" "HMOX1" "FXYD2" "RBP4" "SLC6A12"
[55] "KDELR3" "ITM2B"
In my version of data2, none of the input geneIDs fails to find a mapping. The warning shows that 19.54% of yours fail.
I would have to know much more than I ever will to be able to guess whether this is due to the nature of the beast (the inputIDs) or whether some of them may be malformed, mis-transcribed or simply outside of the reference database.
In sum, there doesn't appear to be anything wrong with your code; the trouble springs from your data. In your position, I would take random samples of, say 25% without replacement and give the result to your function in place of data2 and see how often you get the warning and whether the percentages vary. If you consistently find around 20% fail to map rate, you can be confident that the gene IDs are scattered throughout, and the challenge will be to identify them.
Lets say you take six samples, a,b,c,d,e,f`` that produce18.04, 19.25, 18.97, 19.01 and 21.2` in the warnings.
Do setdiff on each pair to find the unions, a',b',c',d',e',f' and run those through the function and note the differences in results. Proceeding that way will help you narrow down the possible offenders, subset them out of data2 and repeat, eventually to allow you to build a list of known problematic geneIDs to ether be censored, corrected or, if this is an expected result for the type of gene set you're working for, to consult the documentation for any functions for parameter tuning on any modeling you're planning.
Thank you for your suggestions, as a beginner of R, there are so many questions:joy:.
There is one more question
ego_MF <- enrichGO(gene = data.df$ENTREZID, universe = names(geneList),OrgDb = org.Mm.eg.db,ont = "MF", pAdjustMethod = "BH",pvalueCutoff = 1,qvalueCutoff = 1,readable = FALSE)
Error in enricher_internal(gene, pvalueCutoff = pvalueCutoff, pAdjustMethod = pAdjustMethod, :
object 'geneList' not found
I don‘t know witch step is wrong
@songh, this is indeed an R question!
There generally are a variable number of steps needed to resolve error messages:
object 'geneList' not found and the calling function:ego_MF <- enrichGO(gene = data.df$ENTREZID, universe = names(geneList),OrgDb = org.Mm.eg.db,ont = "MF", pAdjustMethod = "BH",pvalueCutoff = 1,qvalueCutoff = 1,readable = FALSE)
Read the help page for the function that is returning the error:
??enrichGO
Read the description of what the function is intended to do:
GO Enrichment Analysis of a gene set. Given a vector of genes, this function will return the enrichment GO categories after FDR control
enrichGO(gene, OrgDb, keyType = "ENTREZID", ont = "MF",
pvalueCutoff = 0.05, pAdjustMethod = "BH", universe,
qvalueCutoff = 0.2, minGSSize = 10, maxGSSize = 500,
readable = FALSE, pool = FALSE)
universe
background genes
geneList that you provided is in your working environment: help(names)
> ls()
[1] "data.df" "data2" "de" "eg" "fit" **"geneList"** "ids" "idx_date"
[9] "mbw" "rbw" "rby" "tbw" "x" "yy"
If it isn't, do you need to create one? If so, how?
See how it is used in the example:
> data(geneList, package = "DOSE")
> de <- names(geneList)[1:100]
> yy <- enrichGO(de, 'org.Hs.eg.db', ont="BP", pvalueCutoff=0.01)
While you're there call head(yy) to see if the output contains the information you expected.
Look at how enrichGO was called. There's nothing after the pvalueCutoff argument. Why?
Go back to the description of universe. Are you using background genes?
Look at the first argument in the example de and its definition in the line above. Is that how you defined geneList
Note the following brackets [1:100] which selects the first 100 genes from the geneList defined in the DOSE package. If you wanted to use the second 100, it would be [101:200]
If you defined your own geneList compare it to the 100 genes used in the example
> tmp <- data(geneList, package = "DOSE")
> head(tmp)
> > head(geneList)
4312 8318 10874 55143 55388 991
4.572613 4.514594 4.418218 4.144075 3.876258 3.677857
rm(tmp) # good practice is to remove temporary variables once you have used them
geneList is it in the same form as the example output?The hardest thing about R is that it is a functional language. It's also the easiest thing. Recall from school
f(x) = y
and then think of all the times when x was expanded to
f(a,b,c,d) = y
In R, it's exactly the same
someFunction <- y = f(x) = y
anotherFunction <- f(a,b,c,d) = y
While there are some features in R that are familiar from most program languages in the procedural/imperative style, such as C, C++, Java, Perl, Ruby, Python, such as for loops, they are only a small part. Usually you will be in the functional mode and most of the time debugging errors is like I've described above.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.