Question

EnhancedVolcano plot error: Log2foldchange is not numeric!

2

Entering edit mode

5.0 years ago

michelle.piquet ▴ 60

I'm trying to make a volcano plot with the differential expression file generated from cellranger for single cell analysis.

I uploaded the file into RStudio and ran the following code...

EnhancedVolcano(res, lab = rownames(results), x = 'Log2foldchange', y = 'pvalue', xlim = c(-8,8), title = 'Macrophage', pCutoff = 10e-16, FCcutoff = 1.5, pointSize = 3.0, labSize = 3.0)

and received the following error...

Error in EnhancedVolcano(x, lab = rownames(x), x = "Log2foldchange", y = "pvalue",  : 
  Log2foldchange is not numeric!

I then ran: is.numeric(results$gene.Log2foldchange.pvalue) and got TRUE.

Any guidance as to what to do?

Thanks!

R EnhancedVolcano • 7.5k views

ADD COMMENT • link 5.0 years ago by michelle.piquet ▴ 60

1

Entering edit mode

Output of colnames(res)? Probably your colnames do not match what you've given as xand y.

ADD REPLY • link 5.0 years ago by ATpoint 85k

0

Entering edit mode

colnames(results) [1] "gene.Log2foldchange.pvalue"

ADD REPLY • link 5.0 years ago by michelle.piquet ▴ 60

1

Entering edit mode

Provide example data or paste the output of str(res).

ADD REPLY • link 5.0 years ago by zx8754 12k

0

Entering edit mode

'data.frame':   31328 obs. of  62 variables:
 $ Feature.ID                 : Factor w/ 31328 levels "ENSG00000000003.14",..: 22057 20888 27958 22670 22625 30594 19475 21679 21805 19878 ...
 $ Feature.Name               : Factor w/ 31096 levels "1-Dec","1-Mar",..: 20867 4796 20866 20865 20868 8794 2191 24343 24342 24626 ...
 $ Cluster.1.Mean.Counts      : num  0.000961 0.00032 0.001281 0.00032 0.00032 ...
 $ Cluster.1.Log2.fold.change : num  0.292 1.955 -2.009 -2.482 3.54 ...
 $ Cluster.1.Adjusted.p.value : num  1 0.9102 0.0467 0.0809 0.6103 ...
 $ Cluster.2.Mean.Counts      : num  0.000902 0.001805 0.006316 0.004511 0 ...
 $ Cluster.2.Log2.fold.change : num  0.793 4.77 0.352 0.839 3.185 ...
 $ Cluster.2.Adjusted.p.value : num  1 0.154 1 0.738 1 ...
 $ Cluster.3.Mean.Counts      : num  0.00181 0 0.01905 0.00726 0 ...
 $ Cluster.3.Log2.fold.change : num  1.46 2.19 2.01 1.5 3.19 ...
 $ Cluster.3.Adjusted.p.value : num  0.7578 1 0.0326 0.3332 1 ...
 $ Cluster.4.Mean.Counts      : num  0.000642 0 0.002568 0.004493 0 ...
 $ Cluster.4.Log2.fold.change : num  0.269 1.662 -0.887 0.779 2.662 ...
 $ Cluster.4.Adjusted.p.value : num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.5.Mean.Counts      : num  0 0 0.02408 0.00219 0 ...
 $ Cluster.5.Log2.fold.change : num  1.048 3.508 2.311 0.485 4.508 ...
 $ Cluster.5.Adjusted.p.value : num  1 1 0.21 1 1 ...
 $ Cluster.6.Mean.Counts      : num  0 0 0.00833 0.00238 0 ...
 $ Cluster.6.Log2.fold.change : num  0.143 2.602 0.769 0.187 3.602 ...
 $ Cluster.6.Adjusted.p.value : num  1 1 1 1 1 ...
 $ Cluster.7.Mean.Counts      : num  0 0 0 0 0 ...
 $ Cluster.7.Log2.fold.change : num  1.6494 4.1088 -0.8101 0.0644 5.1088 ...
 $ Cluster.7.Adjusted.p.value : num  1 1 1 1 1 ...
 $ Cluster.8.Mean.Counts      : num  0.00079 0 0 0.00079 0 ...
 $ Cluster.8.Log2.fold.change : num  0.59 1.98 -2.94 -1.04 2.98 ...
 $ Cluster.8.Adjusted.p.value : num  1 1 0.967 1 1 ...
 $ Cluster.9.Mean.Counts      : num  0.000565 0 0.014123 0.006214 0 ...
 $ Cluster.9.Log2.fold.change : num  0.0694 1.4617 1.5772 1.2653 2.4617 ...
 $ Cluster.9.Adjusted.p.value : num  1 1 0.0436 0.4054 1 ...
 $ Cluster.10.Mean.Counts     : num  0.00224 0 0.01119 0.00895 0 ...
 $ Cluster.10.Log2.fold.change: num  1.77 2.51 1.18 1.82 3.51 ...
 $ Cluster.10.Adjusted.p.value: num  1 1 1 0.966 1 ...
 $ Cluster.11.Mean.Counts     : num  0.00137 0 0.00137 0.00137 0 ...
 $ Cluster.11.Log2.fold.change: num  0.802 0.65 -1.899 -0.983 1.65 ...
 $ Cluster.11.Adjusted.p.value: num  1 1 0.066 0.613 1 ...
 $ Cluster.12.Mean.Counts     : num  0.00303 0 0.00606 0.00454 0 ...
 $ Cluster.12.Log2.fold.change: num  2.226 2.963 0.415 0.986 3.963 ...
 $ Cluster.12.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.13.Mean.Counts     : num  0 0 0.00557 0.00223 0 ...
 $ Cluster.13.Log2.fold.change: num  0.0439 2.5034 0.2303 0.0883 3.5034 ...
 $ Cluster.13.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.14.Mean.Counts     : num  0 0 0.0048 0.0096 0 ...
 $ Cluster.14.Log2.fold.change: num  1.183 3.643 0.333 2.011 4.643 ...
 $ Cluster.14.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.15.Mean.Counts     : num  0 0 0 0 0 ...
 $ Cluster.15.Log2.fold.change: num  2.5366 4.996 0.0772 0.9516 5.996 ...
 $ Cluster.15.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.16.Mean.Counts     : num  0 0 0.005783 0.000826 0 ...
 $ Cluster.16.Log2.fold.change: num  -0.409 2.051 0.218 -0.972 3.051 ...
 $ Cluster.16.Adjusted.p.value: num  1 1 1 1 1 ...
 $ Cluster.17.Mean.Counts     : num  0.00228 0 0.00342 0.00456 0 ...
 $ Cluster.17.Log2.fold.change: num  1.802 2.539 -0.344 0.906 3.539 ...
 $ Cluster.17.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.18.Mean.Counts     : num  0 0 0 0.00695 0 ...
 $ Cluster.18.Log2.fold.change: num  1.725 4.184 -0.735 1.769 5.184 ...
 $ Cluster.18.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.19.Mean.Counts     : num  0.001786 0 0.003572 0.000893 0 ...
 $ Cluster.19.Log2.fold.change: num  1.432 2.169 -0.379 -0.853 3.169 ...
 $ Cluster.19.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...
 $ Cluster.20.Mean.Counts     : num  0 0 0.00658 0 0 ...
 $ Cluster.20.Log2.fold.change: num  1.6449 4.1043 0.7945 0.0599 5.1043 ...
 $ Cluster.20.Adjusted.p.value: num  1 1 1 1 1 1 1 1 1 1 ...

ADD REPLY • link updated 5.0 years ago by zx8754 12k • written 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

So where are the "Log2foldchange" and "pvalue"columns?

ADD REPLY • link 5.0 years ago by zx8754 12k

0

Entering edit mode

I started again from the beginning.

I loaded my data with:

gene_list <- read.table("differential_expression_macrophage.csv", header = T, sep = ",")
31328 obs of 62 variable.

Then tried the volcano plot code and I'm still getting the same error.

is.numeric(res$Cluster.1.Log2.fold.change) [1] TRUE 
Error in EnhancedVolcano(x, lab = rownames(x), x = "Cluster 1 Log2 fold change", y = "Cluster 1 Adjusted pvalue", : Cluster.1.Log2>fold.change is not numeric!

ADD REPLY • link updated 5.0 years ago by zx8754 12k • written 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

Please check your column names:

res$Cluster.1.Log2.fold.change

vs how you supplying to the function:

x = "Cluster 1 Log2 fold change"

ADD REPLY • link 5.0 years ago by zx8754 12k

0

Entering edit mode

Indeed, for your data, this should work:

EnhancedVolcano(
  res2,
  lab = as.character(res2$Feature.Name),
  x = 'Cluster.1.Log2.fold.change',
  y = 'Cluster.1.Adjusted.p.value')

Thanks to everyone who has been providing feedback here...

ADD REPLY • link 5.0 years ago by Kevin Blighe 88k

0

Entering edit mode

Why are you checking with is.numeric(res$gene.Log2foldchange.pvalue) ? Instead, you should be checking is.numeric(res$Log2foldchange)!!

ADD REPLY • link 5.0 years ago by Hyper_Odin ▴ 320

0

Entering edit mode

 is.numeric(results$Log2foldchange)
[1] FALSE

It is indeed FALSE, how do I change it?

ADD REPLY • link updated 5.0 years ago by Ram 44k • written 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

It is also FALSE when then column does not exist. Are you sure your columns that store p and FC are named Log2foldchange and pvalue?

ADD REPLY • link 5.0 years ago by ATpoint 85k

0

Entering edit mode

Yes, they are stored the way they are named.

ADD REPLY • link 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

Try: res$Log2foldchange <- as.numeric(res$Log2foldchange)

ADD REPLY • link 5.0 years ago by zx8754 12k

0

Entering edit mode

I did and ran the same code as before and get the same error.

ADD REPLY • link 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

I upload a new data set and and now I get TRUE still get the same error.

 is.numeric(res$Cluster.1.Log2.fold.change)
[1] TRUE
Error in EnhancedVolcano(x, lab = rownames(x), x = "Cluster 1 Log2 fold change", y = "Cluster 1 Adjusted pvalue",  : 
  Log2foldchange is not numeric!

ADD REPLY • link updated 5.0 years ago by Ram 44k • written 5.0 years ago by michelle.piquet ▴ 60

0

Entering edit mode

Are you sure the table is in "x" ??

Because earlier you stored it in res!!

EnhancedVolcano(x, lab = rownames(x), x = "Cluster 1 Log2 fold change", y = "Cluster 1 Adjusted pvalue",

It should be

EnhancedVolcano(res, lab = rownames(x), x = "Cluster.1.Log2.fold.change", y = "Cluster.1.Adjusted.pvalue"`,

right??

Check the x and y names if they are matching to the res table!!

ADD REPLY • link updated 5.0 years ago by Ram 44k • written 5.0 years ago by Hyper_Odin ▴ 320

0

Entering edit mode

I ran this code and it works but doesn't show the gene names.

EnhancedVolcano(gene_list, lab = rownames(gene_list), x = 'Cluster.1.Log2.fold.change', y = 'Cluster.1.Adjusted.p.value', xlim = c(-5,8), aes(label = Feature.Name))

Any suggestions?

ADD REPLY • link updated 5.0 years ago by GenoMax 147k • written 5.0 years ago by michelle.piquet ▴ 60

1

Entering edit mode

That is because the rownames of your data-frame, called gene_list, are not set. I can see this via the output of

str(gene_list)

Also, if you type rownames(gene_list) and press return / enter, you will not see any gene names.

You need to use:

EnhancedVolcano(
  res2,
  lab = as.character(res2$Feature.Name),
  x = 'Cluster.1.Log2.fold.change',
  y = 'Cluster.1.Adjusted.p.value',
  xlim = c(-5,8))

Why have you included the aes() function, by the way? - that will not do anything.

---------------------------

Can you please go through the entire vignette for EnhancedVolcano so that you can begin to understand how to use the package. I developed the package and wrote the vignette myself: http://bioconductor.org/packages/release/bioc/vignettes/EnhancedVolcano/inst/doc/EnhancedVolcano.html

ADD REPLY • link 5.0 years ago by Kevin Blighe 88k