dplyr mutate new columns

Question

[1769mkc]

 gene    HSC_7256  HSC_6792  HSC_7653 HSC  SU048_Blast SU209_Blast

APOBEC3A   -0.9621548 0.000000 -1.235946  0.000000 0.8818193  -0.9603217
ATAD2B    5.6414714  5.402059 5.79380  5.755443  6.1243134  5.4867616

I m trying to get mean of these multiple columns like HSC one column and Blast another column where mean of each sample would be calculated.

Im doing this

j <- HSC_Blast_EPI_factor_gene %>% rowwise() %>% 
  mutate(HSC =mean(c(HSC_7256,HSC_6792,HSC_7653,HSC),na.rm = TRUE))

then I get a new column , but Im not able to pipe the same then when I try to do for the sample after HSC

When I doing this

j <- HSC_Blast_EPI_factor_gene %>% rowwise() %>% 
  mutate(HSC =mean(c(HSC_7256,HSC_6792,HSC_7653,HSC),
  mutate(LSC =mean(c(SU048_Blasts,SU209_Blasts))),na.rm = TRUE))

I get this error

> Error in mutate_impl(.data, dots) :    Evaluation error: argument
> ".data" is missing, with no default.

what am i doing wrong ?any suggestion or help would be appreciated

[andrew.j.skelton73]

Why not melt the data frame (or gather) instead of moving onto the pain of rowwise operations? rowwise operations never really worked for me.

[cpad0112]

Your code has two issues:

1. Column names mismatch between code and data frame (SU048_Blast vs SU048_Blasts)
2. No need to use mutate function twice.

Your code should be:

j <- HSC_Blast_EPI_factor_gene %>%  rowwise() %>% mutate(HSC = mean(c(HSC_7256,HSC_6792,HSC_7653,HSC)), LSC = mean(c(SU048_Blast,SU209_Blast)))

[1769mkc]

thank you for rectifying me

[cpad0112]

Is it working now?

[1769mkc]

its working perfectly fine ..instead of comment if you could have answered then i would have vote up all the answers

[cpad0112]

This is not answer to your question. However there is another way to do the same in R:

> test
      gene   HSC_7256 HSC_6792  HSC_7653      HSC SU048_Blast SU209_Blast
1 APOBEC3A -0.9621548 0.000000 -1.235946 0.000000   0.8818193  -0.9603217
2   ATAD2B  5.6414714 5.402059  5.793800 5.755443   6.1243134   5.4867616

output:

> cbind(test, HSC=apply(test[,grep ("hsc", ignore.case = T, names(test))],1,mean), SU=apply(test[,grep ("su", ignore.case = T, names(test))],1,mean))
      gene   HSC_7256 HSC_6792  HSC_7653      HSC SU048_Blast SU209_Blast        HSC         SU
1 APOBEC3A -0.9621548 0.000000 -1.235946 0.000000   0.8818193  -0.9603217 -0.5495252 -0.0392512
2   ATAD2B  5.6414714 5.402059  5.793800 5.755443   6.1243134   5.4867616  5.6481933  5.8055375

[1769mkc]

okay I will try your way ,i was just trying to see what is going wrong its a pretty simple operation...isn't it ..

[cpad0112]

dplyr solution without rowwise function:

> test
      gene   HSC_7256 HSC_6792  HSC_7653      HSC SU048_Blast SU209_Blast
1 APOBEC3A -0.9621548 0.000000 -1.235946 0.000000   0.8818193  -0.9603217
2   ATAD2B  5.6414714 5.402059  5.793800 5.755443   6.1243134   5.4867616

Command:

> test %>% mutate(HSC_mean = rowMeans(.[grep("HSC", names(.))]), SU_mean = rowMeans(.[grep("SU", names(.))]))

Output:

      gene   HSC_7256 HSC_6792  HSC_7653      HSC SU048_Blast SU209_Blast   HSC_mean    SU_mean
1 APOBEC3A -0.9621548 0.000000 -1.235946 0.000000   0.8818193  -0.9603217 -0.5495252 -0.0392512
2   ATAD2B  5.6414714 5.402059  5.793800 5.755443   6.1243134   5.4867616  5.6481933  5.8055375

Well, rowwise and two mutate functions are not necessary (in OP code). I guess for the second mutate function, it was looking for data to operate and could not find (IANP/S).