Subset and loop to export the pdf files

mtoufiq · February 26, 2022, 7:41pm

Hi,

I am working with dataframes in R one containing the quantitative data and other one with sample metadata and trying to plot a heatmap using library(ComplexHeatmap) package. I am currently analyzing by subsetting based on each "Group" individually and exporting the pdf plot. On my large datasets, this process becomes very tedious and cumbersome. I am exploring a way to loop to subset the data based on the "Group A, B, and C" from the sample metadata and generate separate plots in the form of the pdf file. I have provided the input data and expected results below. Please advise on how to solve this.

Thank you,
Toufiq

## Quantitative input data
dput(my_matrix_Z)
structure(c(0.932604654502187, -0.435436614072738, 0.477426784581654, 
            -1.46076832746542, -1.31037812634022, -1.01424032341432, -0.0404822682927365, 
            -0.539937618199118, 1.38721958124815, 0.707658540034404, 0.892420147849494, 
            1.03812906674728, -0.0358377647341346, -1.43311295375476, -0.940550471704507, 
            0.0203370331247248, -0.871773522040406, 1.04199433191731, 0.270457448714327, 
            0.581830150089182, -1.38089368299436, -0.60547605837307, -1.88380551789616, 
            -1.53172474098731, 0.772474453928738, -0.227861010731477, -0.656505388793654, 
            1.09553965503928, 0.0346508158080517, 1.61146228986793, 0.5749024862802, 
            0.78190001284444, 0.471933590622001, 0.355011798526725, 1.30613677926137, 
            0.014694768806926), .Dim = c(4L, 9L), .Dimnames = list(c("FFBB1", 
                                                                     "CCVV2", "XBPA3", "ALOT4"), c("Sample_1", "Sample_2", "Sample_3", 
                                                                                                   "Treat4", "Treat5", "Treat6", "Treat_7", "Treat_8", "Treat_9"
                                                                     )))
#>         Sample_1    Sample_2  Sample_3      Treat4     Treat5     Treat6
#> FFBB1  0.9326047 -1.31037813 1.3872196 -0.03583776 -0.8717735 -1.3808937
#> CCVV2 -0.4354366 -1.01424032 0.7076585 -1.43311295  1.0419943 -0.6054761
#> XBPA3  0.4774268 -0.04048227 0.8924201 -0.94055047  0.2704574 -1.8838055
#> ALOT4 -1.4607683 -0.53993762 1.0381291  0.02033703  0.5818302 -1.5317247
#>          Treat_7    Treat_8    Treat_9
#> FFBB1  0.7724745 0.03465082 0.47193359
#> CCVV2 -0.2278610 1.61146229 0.35501180
#> XBPA3 -0.6565054 0.57490249 1.30613678
#> ALOT4  1.0955397 0.78190001 0.01469477

## Sample metadata
dput(Sample_Info)
structure(list(Group = c("A", "A", "A", "B", "B", "B", "C", "C", 
                         "C"), Treatment = c("IL1", "Cyt", "Unt", "IL1", "Cyt", "Unt", 
                                             "IL1", "Cyt", "Unt")), class = "data.frame", row.names = c("Sample_1", 
                                                                                                        "Sample_2", "Sample_3", "Treat4", "Treat5", "Treat6", "Treat_7", 
                                                                                                        "Treat_8", "Treat_9"))
#>          Group Treatment
#> Sample_1     A       IL1
#> Sample_2     A       Cyt
#> Sample_3     A       Unt
#> Treat4       B       IL1
#> Treat5       B       Cyt
#> Treat6       B       Unt
#> Treat_7      C       IL1
#> Treat_8      C       Cyt
#> Treat_9      C       Unt

## Subset the data;
my_matrix_Z_A <- my_matrix_Z[, c(1:3)]
Sample_Info_A <- Sample_Info[c(1:3), ]
my_matrix_Z_B <- my_matrix_Z[, c(4:6)]
Sample_Info_B <- Sample_Info[c(4:6), ]
my_matrix_Z_C <- my_matrix_Z[, c(7:9)]
Sample_Info_C <- Sample_Info[c(7:9), ]

## Column annotations for all "Groups"
library(ComplexHeatmap)

column_ha = HeatmapAnnotation(df = data.frame(Group = Sample_Info$Group,
                                              Treatment = Sample_Info$Treatment),
                              col = list(Group = c("A" = "#DC61C5", "B" = "#9843DA", "C" = "#DDA862"),
                                         Treatment =c("IL1" = "#CBD7E1", "Cyt" = "#D3A9D4",  "Unt" = "#7B9F7F")))

## Column annotations for all Groups A
column_ha_A = HeatmapAnnotation(df = data.frame(Group = Sample_Info_A$Group,
                                              Treatment = Sample_Info_A$Treatment),
                              col = list(Group = c("A" = "#DC61C5"),
                                         Treatment =c("IL1" = "#CBD7E1", "Cyt" = "#D3A9D4",  "Unt" = "#7B9F7F")))

## Export the plot for all Groups
pdf("Plot_for_Group_All.pdf",height = 5, width = 5)
Heatmap(my_matrix_Z,
        col = circlize::colorRamp2(c(-1, 0, 1), c("blue", "white", "red")),
        cluster_columns = F,
        top_annotation = column_ha)
dev.off()

## Export the plot for "Group A"
pdf("Plot_for_Group_All.pdf",height = 5, width = 5)
Heatmap(my_matrix_Z_A,
        col = circlize::colorRamp2(c(-1, 0, 1), c("blue", "white", "red")),
        cluster_columns = F,
        top_annotation = column_ha_A)
dev.off()

Likewise, 
## Export the plot for "Group B"
## Export the plot for "Group C"

Expected results:

Export plots as individual files

^{Created on 2022-02-26 by the reprex package (v2.0.1)}

pieterjanvc · February 28, 2022, 2:34pm

Hi there,

Thanks for the nice reprex (reproducible example), this really helped building a solution!
Here is my code for looping. I did omit the code for the plot that has all the data in it and focused on generating the different subplots efficiently

library(ComplexHeatmap)
require(circlize)

## Quantitative input data
my_matrix_Z = structure(
  c(
    0.932604654502187,-0.435436614072738, 0.477426784581654,-1.46076832746542,
    -1.31037812634022,-1.01424032341432,-0.0404822682927365,-0.539937618199118,
    1.38721958124815, 0.707658540034404,0.892420147849494,1.03812906674728,
    -0.0358377647341346,-1.43311295375476,-0.940550471704507,0.0203370331247248,
    -0.871773522040406,1.04199433191731,0.270457448714327,   
    0.581830150089182,-1.38089368299436,-0.60547605837307,    
    -1.88380551789616,-1.53172474098731,0.772474453928738,   
    -0.227861010731477,-0.656505388793654,    1.09553965503928,
    0.0346508158080517, 1.61146228986793,0.5749024862802, 0.78190001284444,
    0.471933590622001,0.355011798526725,1.30613677926137, 0.014694768806926  ),
  .Dim = c(4L, 9L),
  .Dimnames = list(
    c("FFBB1","CCVV2", "XBPA3", "ALOT4"),
    c("Sample_1","Sample_2","Sample_3","Treat4","Treat5",
      "Treat6","Treat_7","Treat_8", "Treat_9")
  )
)

## Sample metadata
Sample_Info = structure(
  list(
    Group = c("A", "A", "A", "B", "B", "B", "C", "C","C"),
    Treatment = c("IL1", "Cyt", "Unt", "IL1", "Cyt", "Unt",
                  "IL1", "Cyt", "Unt")
  ),
  class = "data.frame",
  row.names = c("Sample_1", "Sample_2", "Sample_3",
    "Treat4", "Treat5", "Treat6", "Treat_7","Treat_8","Treat_9"
  )
)


### CHANGES START HERE ####

#Add extra Name column (rownames) for easier subsetting
Sample_Info$Name = rownames(Sample_Info)
#Get a named vector group colours (can be generated too depending on data)
colours = c("A" = "#DC61C5", "B" = "#9843DA", "C" = "#DDA862")

#Loop over all groups
for(group in unique(Sample_Info$Group)){
  
  #Subset the data
  my_matrix = my_matrix_Z[,Sample_Info$Name[Sample_Info$Group == group]]
  
  ## Column annotations for group
  column_ha = HeatmapAnnotation(
    df = data.frame(Group = Sample_Info$Group[Sample_Info$Group == group],
                    Treatment = Sample_Info$Treatment[Sample_Info$Group == group]),
    col = list(
      Group = colours[names(colours) == group],
      Treatment = c(
        "IL1" = "#CBD7E1",
        "Cyt" = "#D3A9D4",
        "Unt" = "#7B9F7F"
      )
    )
  )
  
  #Plot the group
  pdf(sprintf("Plot_for_Group_%s.pdf", group), height = 5, width = 5)
  #important to wrap in the plot() function for correct saving of the data
  plot(Heatmap(
    my_matrix,
    col = circlize::colorRamp2(c(-1, 0, 1), c("blue", "white", "red")),
    cluster_columns = F,
    top_annotation = column_ha
  ))
  dev.off()
}

Hope this helps,
PJ

mtoufiq · March 1, 2022, 10:17am

@pieterjanvc,

thank you very much for the detailed response. I am actually running the same code that you provided, however, this shows an error message as indicated below. Do you know what is the reason for this error? Thank you.

Error in as.double(y) :
cannot coerce type 'S4' to vector of type 'double'

pieterjanvc · March 1, 2022, 12:53pm

Hi,

I ran the code again and it's working well on my machine. Do you have both libraries loaded and the latest version installed? It seems a package might be missing or not updated. Guessing from the error maybe the S4Vectors package. Try and install ComplexHeatmap again (which contains that package) and see if it works.

PJ

mtoufiq · March 1, 2022, 2:14pm

@pieterjanvc,

oops, I just recalled the past situation now. I had updated the ComplexHeatmap library to the latest version and some of the heatmap code did not work in the R Shiny app. Hence, I had to re-install the previous version of the ComplexHeatmap library again to make it work. In case, if I install the latest version, then again I might face many issues while re-deploying Shiny apps and almost all apps are using this package. Is there a workaround to fix this issue? Thank you very much.

pieterjanvc · March 1, 2022, 5:14pm

Hi,

Which version of R are you using, and where does the error occur? I assume it will be in the last part where the plot() function is called.

Also, try and see if the issue can be solved by replacing the plot() by either
graphics::plot() or base::plot()

PJ

mtoufiq · March 1, 2022, 8:07pm

@pieterjanvc,

I am using R version 4.1.1 . Yes, the error occurs at last part where the plot() function is used. In addition tried these graphics::plot() or base::plot() but still, unsuccessful.

sessionInfo()
R version 4.1.1 (2021-08-10)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Catalina 10.15.6

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] grid      stats     graphics  grDevices utils     datasets 
[7] methods   base     

other attached packages:
[1] circlize_0.4.13      ComplexHeatmap_2.8.0

loaded via a namespace (and not attached):
 [1] IRanges_2.26.0      matrixStats_0.60.1  codetools_0.2-18   
 [4] png_0.1-7           crayon_1.4.1        digest_0.6.28      
 [7] foreach_1.5.1       stats4_4.1.1        GlobalOptions_0.1.2
[10] doParallel_1.0.16   S4Vectors_0.30.2    GetoptLong_1.0.5   
[13] RColorBrewer_1.1-2  rjson_0.2.20        Cairo_1.5-12.2     
[16] iterators_1.0.13    tools_4.1.1         tinytex_0.34       
[19] xfun_0.27           parallel_4.1.1      compiler_4.1.1     
[22] clue_0.3-60         BiocGenerics_0.38.0 colorspace_2.0-2   
[25] cluster_2.1.2       shape_1.4.6

pieterjanvc · March 1, 2022, 10:31pm

HI,

Try and update the S4Vectors package and see if that helps

PJ

mtoufiq · March 2, 2022, 7:54am

@pieterjanvc, I tried updating the S4Vector package as well, still unsuccessful.

Lastly, I had to re-install latest the Complexheatmap package using the below source:

library(devtools)
install_github("jokergoo/ComplexHeatmap", force=TRUE)

Thank you for the continuous support. I was able to get the plots and this question is resolved now.

system · March 9, 2022, 7:55am

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.