Pheatmap/Complexheatmap: making a continuous color scale with NAs
1
0
Entering edit mode
@d1feef57
Last seen 2.2 years ago
Rijnsburg

Hello everyone, hope you are doing well:)

I have already posted this question in Biostars, but didn't get an answer there, so I decided to post it here.

I have a question regarding making a heatmap to visualize significance AND presense of pathways across different conditions. I already did pathway enrichment analysis using gProfiler. I imported the results and manipulated them to make the names of the pathways as rownames, adjusted p values in rows and colnames for the conditions like so:

    kegg_react_wp[1:3,] # showing only first 3 pathways

                              Condition1    Condition2        Condition3     Condition4     Condition5   Condition6
                      Pathway1       0.0003            0.3225              0.0540            0.0003      0.01         0.07
                      Pathway2       0.03             0.003225              0.0540            0.0003          0.01         0.07
                      Pathway3       0.703            0.3225              0.0540            0.0003            0.01         0.07

Because some pathways have an adjusted p value >0.05, I want to make these pathways as NA , which indicates that this pathway was NOT enriched in this condition. So I used the following code to do that as well as to -log10 the pathways with adjusted p value<0.05 so that most significant pathways have larger values:

     Mooi_functie= function(x){
      x = ifelse(x>0.05,NA,-log10(x))
      return(x)
    }

Now apply the function to my data

    pathways_clean<-kegg_react_wp%>%
    mutate_if(is.numeric,Mooi_functie) # this will make cells NA if their adjusted p value larger than 0.05

Now making the heatmap(using pheatmap or complexheatmap):

     pheatmap(pathways_clean) # pheatmap package
     Heatmap(pathways_clean)# complexheatmap package

I get the following error

     Error in hclust(d, method = method) : 
      NA/NaN/Inf in foreign function call (arg 10) # from pheatmap
     Error in hclust(get_dist(submat, distance), method = method) : 
     NA/NaN/Inf in foreign function call (arg 10) # from complexheatmap

A similar question has been posted here about the same error, but the purpose for that question, was different than mine. I want to KEEP NAs(unlike the already posted question) so that I can color them differently to indicate that these pathways are NOT present in this condition. it seems that this error has to do with the clustering of the rows, so if I set cluster_rows=F, it works, but I want to cluster the rows and scale by row to see in which condition the significance is larger. I understand that there are some rows which are basiaclly all NAs except for one or two cells and this seems problematic for making the heatmaps.

I found in the internet a nice trick to add a column with any value , and it worked

 pathways_clean$fake_column<-1

However, now It worked, but I am left with a fake column. it's not nice to have this column in the heatmap as you can see below.

the code I used to generate the heatmap:

  pheatmap(pathways_clean,cluster_rows = T,na_col = "white",border_color = "white",
         annotation_row =meta_kegg_wp_reac,cellwidth = 35,fontsize = 8,angle_col = 45,
         scale = "row")

enter image description here

My questions are:

1) How to solve the issue of clustering with NAs without making a fake column? Or perhaps there is another way to visualize this without using NAs?(I tried replacing the NAs with zeros, but then the color scaling gets messy and cannot tell heads from tails).

2) Does it make sense to scale( scale="row") the -log10 of adjusted p values? Because I find it difficult to make sense of z-scores of -log10 of adjusted p values, which is what the legend in the heatmap represents.

3) Do you think that there might be a better way to visualize these results in my case?

Thank you very much in advance for your help!

**# My session info**
sessionInfo( )


R version 4.0.3 (2020-10-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)

Matrix products: default

locale:
[1] LC_COLLATE=Dutch_Netherlands.1252  LC_CTYPE=Dutch_Netherlands.1252    LC_MONETARY=Dutch_Netherlands.1252
[4] LC_NUMERIC=C                       LC_TIME=Dutch_Netherlands.1252    

attached base packages:
 [1] stats4    grid      parallel  stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] org.Hs.eg.db_3.11.4    AnnotationDbi_1.50.3   IRanges_2.22.2         S4Vectors_0.26.1       clusterProfiler_3.16.1
 [6] treemap_2.4-2          gridBase_0.4-7         eulerr_6.1.0           ggupset_0.3.0          ComplexUpset_1.1.0    
[11] wesanderson_0.3.6      viridis_0.5.1          viridisLite_0.3.0      networkD3_0.4          stringi_1.5.3         
[16] tidyr_1.1.2            quantmod_0.4.18        TTR_0.24.2             xts_0.12.1             zoo_1.8-8             
[21] UpSetR_1.4.0           nVennR_0.2.3           VennDiagram_1.6.20     futile.logger_1.4.3    gridExtra_2.3         
[26] EnhancedVolcano_1.6.0  writexl_1.3.1          readr_1.4.0            scales_1.1.1           ggrepel_0.9.1         
[31] esc_0.5.1              sjstats_0.18.1         sjmisc_2.8.6           sjlabelled_1.1.7       ggeffects_1.0.1       
[36] strengejacke_0.5.0     sjPlot_2.8.7           pheatmap_1.0.12        cluster_2.1.1          digest_0.6.27         
[41] circlize_0.4.12        RColorBrewer_1.1-2     ComplexHeatmap_2.4.3   ggpubr_0.4.0           ggplot2_3.3.3         
[46] ggsci_2.9              limma_3.44.3           Biobase_2.48.0         BiocGenerics_0.34.0    data.table_1.13.6     
[51] plyr_1.8.6             readxl_1.3.1           dplyr_1.0.4            readstata13_0.9.2     

loaded via a namespace (and not attached):
  [1] estimability_1.3     coda_0.19-4          bit64_4.0.5          knitr_1.31           multcomp_1.4-16     
  [6] generics_0.1.0       callr_3.5.1          cowplot_1.1.1        lambda.r_1.2.4       TH.data_1.0-10      
 [11] usethis_2.0.1        RSQLite_2.2.3        europepmc_0.4        listviewer_3.0.0     bit_4.0.4           
 [16] enrichplot_1.8.1     httpuv_1.5.5         xml2_1.3.2           assertthat_0.2.1     xfun_0.21           
 [21] hms_1.0.0            promises_1.2.0.1     evaluate_0.14        fansi_0.4.2          progress_1.2.2      
 [26] igraph_1.2.6         DBI_1.1.1            htmlwidgets_1.5.3    purrr_0.3.4          ellipsis_0.3.1      
 [31] selectr_0.4-2        backports_1.2.1      insight_0.13.1       grImport2_0.2-0      vctrs_0.3.6         
 [36] remotes_2.2.0        abind_1.4-5          cachem_1.0.3         withr_2.4.1          ggforce_0.3.2       
 [41] triebeard_0.3.0      emmeans_1.5.4        prettyunits_1.1.1    DOSE_3.14.0          crayon_1.4.1        
 [46] pkgconfig_2.0.3      labeling_0.4.2       tweenr_1.0.1         nlme_3.1-152         pkgload_1.1.0       
 [51] devtools_2.3.2       rlang_0.4.10         lifecycle_1.0.0      sandwich_3.0-0       downloader_0.4      
 [56] modelr_0.1.8         cellranger_1.1.0     rsvg_2.1             rprojroot_2.0.2      polyclip_1.10-0     
 [61] Matrix_1.2-18        urltools_1.7.3       carData_3.0-4        boot_1.3-27          base64enc_0.1-3     
 [66] ggridges_0.5.3       GlobalOptions_0.1.2  processx_3.4.5       png_0.1-7            rjson_0.2.20        
 [71] parameters_0.12.0    blob_1.2.1           shape_1.4.5          stringr_1.4.0        qvalue_2.20.0       
 [76] jpeg_0.1-8.1         rstatix_0.7.0        gridGraphics_0.5-1   ggsignif_0.6.1       memoise_2.0.0       
 [81] magrittr_2.0.1       compiler_4.0.3       scatterpie_0.1.5     clue_0.3-58          lme4_1.1-26         
 [86] snakecase_0.11.0     cli_2.3.1            patchwork_1.1.1      ps_1.5.0             formatR_1.7         
 [91] MASS_7.3-53.1        tidyselect_1.1.0     forcats_0.5.1        yaml_2.2.1           GOSemSim_2.14.2     
 [96] fastmatch_1.1-0      tools_4.0.3          rio_0.5.16           rstudioapi_0.13      foreign_0.8-81      
[101] farver_2.0.3         ggraph_2.0.4         rvcheck_0.1.8        BiocManager_1.30.10  shiny_1.6.0         
[106] Rcpp_1.0.6           car_3.0-10           broom_0.7.5          later_1.1.0.1        performance_0.7.0   
[111] httr_1.4.2           effectsize_0.4.3     colorspace_2.0-0     polylabelr_0.2.0     rvest_0.3.6         
[116] XML_3.99-0.5         fs_1.5.0             splines_4.0.3        statmod_1.4.35       graphlayouts_0.7.1  
[121] ggplotify_0.0.5      sessioninfo_1.1.1    xtable_1.8-4         jsonlite_1.7.2       nloptr_1.2.2.2      
[126] futile.options_1.0.1 tidygraph_1.2.0      testthat_3.0.2       R6_2.5.0             mime_0.10           
[131] pillar_1.5.0         htmltools_0.5.1.1    glue_1.4.2           fastmap_1.1.0        minqa_1.2.4         
[136] BiocParallel_1.22.0  codetools_0.2-16     fgsea_1.14.0         pkgbuild_1.2.0       mvtnorm_1.1-1       
[141] utf8_1.1.4           lattice_0.20-41      tibble_3.0.6         curl_4.3             zip_2.1.1           
[146] GO.db_3.11.4         openxlsx_4.2.3       survival_3.2-7       rmarkdown_2.7        desc_1.2.0          
[151] munsell_0.5.0        DO.db_2.9            GetoptLong_1.0.5     haven_2.3.1          reshape2_1.4.4      
[156] gtable_0.3.0         bayestestR_0.8.2
R Pathways pheatmap limma • 6.0k views
ADD COMMENT
0
Entering edit mode
SamGG ▴ 360
@samgg-6428
Last seen 19 days ago
France/Marseille/Inserm

Hi,

As hclust (in fact dist) can't compute with NA and pheatmap can't scale row with only 1 or 2 values,

solve yourself these 2 independent questions,

then provide pheatmap with the resulting clustering of the rows and the resulting matrix scaled by row.

A "two cents" from a liker of pheatmap.

ADD COMMENT
0
Entering edit mode

You can get an edge over the web with just visit the website and service that will also help you in find an online Quran Teacher website or you can Learners Quran Academy

ADD REPLY
0
Entering edit mode

Thank you sir for your comment. I thought of ignoring clustering by setting cluster.row=F. However, I am still faced with the problem of the colors. Because I have some -log10(fdr) of 30 and others around 5, the coloring of the heatmap looks not pretty. and you cannot really see contrasts. I tried making my own breaks, but it still looks very skewed towards high values without scaling. Any thoughts on how this can be done?

ADD REPLY
0
Entering edit mode

I would replace pvalue < 0.05 with -log10(0.05) instead of NA. NA makes those pvalues not taken into account, but those pvalues are known. They should not be ignored as if the other pvalues of the row are the only ones that allow a row to be compared another row. Whatever you choose, here how to your own clustering of rows:

row_clustering <- hclust(dist(pathways_clean))

Do your own scaling of -log10(pvalues). If current dynamic range is to big, compress it with one or two sqrt transformations.

pathways_clean <- sqrt(pathways_clean)

If you kept the fake column for some reasons, remove it before display, then

pheatmap(pathways_clean,cluster_rows = row_clustering, na_col = "white", border_color = "white",
         annotation_row =meta_kegg_wp_reac,cellwidth = 35,fontsize = 8,angle_col = 45)  # scale is none by default

Once you are OK with the transformation, you can try to play with breaks to associate the right value with a color. I usually don't take time to carry this out, as my goal of using pheatmap is only exploring relationships.

Of course everyone has to understand how to read a heatmap, ie link the color cells and the color scale, and on which values the trees were based. Are color encoded values are on an absolute scale or relative to each row or column? Are the trees of rows and/or columns based on the absolute values or relative values?

Think about the way you want to interpret the pvalues in your case...

ADD REPLY

Login before adding your answer.

Traffic: 449 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6