Add absolute colours to heat map depending on range

Question

[amy__]

Hi,

I have a heatmap, and I wish to make the points from 0-30 white, 31-50 blue, 51+ dark blue.

However I do not want it to be a gradient, I want solid colours, does anyone have any ideas?

Here is a sample I think:

structure(list(chromosome = c("chr5", "chr5", "chr5", "chr5", 
"chr5", "chr5", "chr5", "chr5", "chr5", "chr5", "chr5", "chr5", 
"chr5", "chr5", "chr5", "chr5", "chr5", "chr5", "chr5", "chr5", 
"chr5", "chr5", "chr5", "chr5", "chr5"), start = 112754872:112754896, 
    end = 112754873:112754897, E132.H.001.1_S96_L008_ = c(171L, 
    172L, 172L, 183L, 184L, 186L, 186L, 187L, 190L, 185L, 192L, 
    197L, 199L, 202L, 202L, 204L, 204L, 205L, 208L, 213L, 211L, 
    214L, 217L, 222L, 221L), E132.H.002.1_S26_L002_ = c(92L, 
    96L, 97L, 97L, 99L, 99L, 102L, 107L, 111L, 112L, 114L, 115L, 
    118L, 118L, 119L, 119L, 120L, 124L, 127L, 125L, 125L, 125L, 
    124L, 125L, 125L)), row.names = c(NA, 25L), class = "data.frame")

From my entire dataframe I am able to produce a heatmap using this code:

df <- exon_noUTR_42_3113_index_samples_combined_APC_coverage

#For heatmap we just need count data, so we need to tell R we only want that information first
Counts <- data.frame(df$E132.H.001.1_S96_L008_, df$E132.H.002.1_S26_L002_, df$E132.H.004.1_S28_L002_, df$E132.H.005_S21_L002_, df$E132.H.006_S22_L002_, df$E132.H.007_S23_L002_, df$E132.H.008_S24_L002_, df$E132.H.009.1_S29_L002_, df$E132.H.010.1_S101_L008_, df$E132.H.011.1_S102_L008_, df$E132.H.012.1_S103_L008_, df$E132.H.013_S25_L002_, df$E132.H.014_S26_L002_, df$E132.H.015_S27_L002_, df$E132.H.016_S28_L002_, df$E132.H.017.1_S104_L008_, df$E132.H.018.1_S105_L008_, df$E132.H.019_S29_L002_, df$E132.H.020_S30_L002_, df$E218.A.001_S6_L001_, df$NG0921001_EKDN210018941.1A_HN2MGDSX2_L2_, df$NG0921002_EKDN210018942.1A_HN2MGDSX2_L3_, df$NG0921003_EKDN210018943.1A_HN2MGDSX2_L3_, df$NG0921004_EKDN210018944.1A_HMYL3DSX2_L4_,  df$NG0921005_EKDN210018945.1A_HN2MGDSX2_L2_, df$NG0921006_EKDN210018946.1A_HN2MGDSX2_L2_, df$NG0921007_EKDN210018947.1A_HN2MGDSX2_L3_, df$NG0921008_EKDN210018948.1A_HN2MGDSX2_L3_, df$NG0921009_EKDN210018949.1A_HN2MGDSX2_L3_, df$NG0921010_EKDN210018950.1A_HN2MGDSX2_L2_, df$NG0921011_EKDN210018951.1A_HN2MGDSX2_L3_, df$NG0921012_EKDN210018952.1A_HN2MGDSX2_L3_, df$NG0921013_EKDN210018953.1A_HN2MGDSX2_L2_, df$NG0921014_EKDN210018954.1A_HN2MGDSX2_L3_, df$NG0921015_EKDN210018955.1A_HN2MGDSX2_L3_, df$NG0921016_EKDN210018956.1A_HN2MGDSX2_L2_, df$NG0921017_EKDN210018957.1A_HN2MGDSX2_L2_, df$NG0921018_EKDN210018958.1A_HN2MGDSX2_L2_, df$NG0921019_EKDN210018959.1A_HN2MGDSX2_L2_, df$NG0921020_EKDN210018960.1A_HN2MGDSX2_L2_, df$NG0921021_EKDN210018961.1A_HN2MGDSX2_L2_, df$NG0921022_EKDN210018962.1A_HMYL3DSX2_L4_, df$NG0921024_EKDN210018964.1A_HMYL3DSX2_L4_, df$NG0921025_EKDN210018965.1A_HN2MGDSX2_L3_, df$NG0921026_EKDN210018966.1A_HN2MGDSX2_L2_, df$NG0921027_EKDN210018967.1A_HN2MGDSX2_L3_, df$NG0921028_EKDN210018968.1A_HN2MGDSX2_L3_, df$NG0921029_EKDN210018969.1A_HN2MGDSX2_L3_, df$NG0921030_EKDN210018970.1A_HN2MGDSX2_L2_, df$NG0921031_EKDN210018971.1A_HN2MGDSX2_L3_, df$NG0921032_EKDN210018972.1A_HN2MGDSX2_L3_, df$NG0921033_EKDN210018973.1A_HN2MGDSX2_L3_, df$NG0921035_EKDN210018974.1A_HMYL3DSX2_L4_, df$X_12M01471_L7_, df$X_12M04992_L6_, df$X_12M10775_L6_, df$X_12M10968_L1_, df$X_12M11537_L6_, df$X_13M01595_L1_, df$X_14M02599_L7_, df$X_14M03641_L1_, df$X_14M06940_L7_, df$X_14M11145_L7_, df$X_15M00729_L1_, df$X_15M11163_L7_, df$E036.H.002_CGTACTAG.GTAAGGAG_L001_, df$E036.H.003_AGGCAGAA.GTAAGGAG_L001_, df$E036.H.004_TCCTGAGC.GTAAGGAG_L001_, df$E036.H.005_GGACTCCT.GTAAGGAG_L001_, df$E036.H.006_TAGGCATG.GTAAGGAG_L001_, df$E036.H.009_GCTACGCT.GTAAGGAG_L001_, df$E036.H.011_AAGAGGCA.GTAAGGAG_L001_, df$E036.H.012_GTAGAGGA.GTAAGGAG_L001_, df$E036.H.013_TAAGGCGA.ACTGCATA_L007_, df$E036.H.014_CGTACTAG.ACTGCATA_L007_, df$E036.H.015_AGGCAGAA.ACTGCATA_L007_, df$E036.H.017_GGACTCCT.ACTGCATA_L007_, df$E036.H.018_TAGGCATG.ACTGCATA_L007_, df$E036.H.019_CTCTCTAC.ACTGCATA_L007_, df$E036.H.021_GCTACGCT.ACTGCATA_L007_, df$E036.H.023_AAGAGGCA.ACTGCATA_L007_, df$E036.H.025_GGACTCCT.GTAAGGAG_L007_, df$E036.H.027_CTCTCTAC.GTAAGGAG_L007_, df$E036.H.028_CAGAGAGG.GTAAGGAG_L007_, df$E036.H.029_TAAGGCGA.ACTGCATA_L006_, df$E036.H.030_CGTACTAG.ACTGCATA_L006_, df$E036.H.031_AGGCAGAA.ACTGCATA_L006_, df$E036.H.034_TAGGCATG.ACTGCATA_L006_, df$E036.H.035_CTCTCTAC.ACTGCATA_L006_, df$E036.H.036_CAGAGAGG.ACTGCATA_L006_, df$E036.H.039_CGAGGCTG.ACTGCATA_L006_, df$E036.H.040_TAAGGCGA.GTAAGGAG_L001_, df$E036.H.042_AGGCAGAA.GTAAGGAG_L001_, df$E036.H.043_TCCTGAGC.GTAAGGAG_L001_, df$E036.H.044_GGACTCCT.GTAAGGAG_L001_, df$E036_H_058_L003_, df$E036_H_059_L007_, df$E036_H_062_L003_, df$E036_H_065_L004_, df$E036_H_068_L007_, df$E036_H_070_L005_, df$E036_H_073_L006_, df$E036_H_075_L007_, df$E036_H_076_L006_, df$E036_H_077_L005_, df$E036_H_078_L006 , row.names =df$start)
Counts <- t(Counts)
Counts1 <- as.matrix(Counts)

png("ALL_Jasmine_practice_heatmap_exons_allsamples.png")

#Plot basic heatmap
pheatmap (Counts1, cluster_rows = F, cluster_cols = F, show_rownames = T, color = colorRampPalette(c("white", "blue", "darkblue"))(50), fontsize = 4)
dev.off()

and I am returned:

As you can see it is a gradient from 0-700 which the colour gradually changing, I don't want this. I would like 0-30 white, 31-50 blue, 51+ dark blue.

The sample names are not as blurry when I save as pdf

Thanks! Amy

[amy__]

[ATpoint]

Please read existing solutions first and then say where you get stuck: https://stackoverflow.com/questions/23938889/how-to-create-discrete-legend-in-pheatmap

[Matthias Zepper]

I have not worked with pheatmap before, but if it accepts discrete input, you could just use cut() on your numerical values to make them discrete:

heatmap_values  <- rnorm(100,35,10)
cut(heatmap_values,breaks=c(0,30,50,Inf), include.lowest=TRUE)