ggplot2

Elegant Graphics for Data Analysis

José Antonio López Gómez

¿Qué es ggplot2?

ggplot2 es un paquete de R para producir visualizaciones de datos. A diferencia de otros paquetes graficos, ggplot2 usa un framework conceptual basado en la gramática de los gráficos. Esto permite crear gráficos de diferentes elementos, en lugar de estar limitados a un conjunto de gráficos predeterminados. A día de hoy es el lider indiscutible en representaciones gráficas en R.

Estructura

Existen 7 elementos que se unen como un conjunto de instrucciones para dibujar un gráfico. Un gráfico presenta al menos tres elementos: Data, mapping, layer.

Estructura

• Data: Los datos que se van a representar, generalmente será un data frame.

• Mappings:Características estéticas (aes) que describiran cómo queremos que los datos se vean en el gráfico (posición, color, relleno, forma, tamaño, etc..).

• Layers: Es la capa que muestra como aparecen los datos. Cada una de ellas tiene 3 partes importantes:

  • Geometría: (geom) Determina como se muestra cada punto, linea, etc..
  • Transformación estadisticas: (stat) Que puede computar nuevas variables de los datos.
  • Ajuste de posición: Determina donde se muestra cada parte de datos.

Instalación

El paquete ggplot2 no se encuentra en R-base, para su uso es necesario descargarlo e instalarlo desde los repositorios de CRAN.

install.packages('ggplot2')

Para poder usarlo hay que cargarlo en el sistema usando la función library()

library('ggplot2')

Diamonds

En el paquete ggplot2 esta incluido el dataset Diamonds que contiene el precio y otros atributos de 54.000 diamantes.

library('ggplot2')
df <- diamonds
summary(df)

     carat               cut        color        clarity          depth      
 Min.   :0.2000   Fair     : 1610   D: 6775   SI1    :13065   Min.   :43.00  
 1st Qu.:0.4000   Good     : 4906   E: 9797   VS2    :12258   1st Qu.:61.00  
 Median :0.7000   Very Good:12082   F: 9542   SI2    : 9194   Median :61.80  
 Mean   :0.7979   Premium  :13791   G:11292   VS1    : 8171   Mean   :61.75  
 3rd Qu.:1.0400   Ideal    :21551   H: 8304   VVS2   : 5066   3rd Qu.:62.50  
 Max.   :5.0100                     I: 5422   VVS1   : 3655   Max.   :79.00  
                                    J: 2808   (Other): 2531                  
     table           price             x                y         
 Min.   :43.00   Min.   :  326   Min.   : 0.000   Min.   : 0.000  
 1st Qu.:56.00   1st Qu.:  950   1st Qu.: 4.710   1st Qu.: 4.720  
 Median :57.00   Median : 2401   Median : 5.700   Median : 5.710  
 Mean   :57.46   Mean   : 3933   Mean   : 5.731   Mean   : 5.735  
 3rd Qu.:59.00   3rd Qu.: 5324   3rd Qu.: 6.540   3rd Qu.: 6.540  
 Max.   :95.00   Max.   :18823   Max.   :10.740   Max.   :58.900  
                                                                  
       z         
 Min.   : 0.000  
 1st Qu.: 2.910  
 Median : 3.530  
 Mean   : 3.539  
 3rd Qu.: 4.040  
 Max.   :31.800  
                 

Scatterplot

Vamos a representar carat (quilates) vs price

ggplot(df, aes(x = carat, y = price)) +
    geom_point()

Scatterplot

Añadimos un tema y cambiamos el color de los puntos

ggplot(df, aes(x = carat, y = price)) +
    geom_point(color = 'steelblue') +
    theme_bw()

Scatterplot

Coloreamos los puntos agrupados por la variable cut. Además movemos la leyenda a la parte superior

ggplot(df, aes(x = carat, y = price)) +
    geom_point(aes(color = cut)) +
    theme_bw() +
    theme(legend.position = 'top')

Scatterplot

ggplot(df, aes(x = carat, y = price)) +
    geom_point(aes(color = cut)) +
    theme_bw() +
    theme(legend.position = 'top') +
    facet_wrap(~ cut)

Boxplot

ggplot(df, aes(y = price)) +
    geom_boxplot()

Boxplot

ggplot(df, aes(x = cut, y = price)) +
    geom_boxplot(color = 'navy') +
    theme_bw()

Boxplot

ggplot(df, aes(x = cut, y = price)) +
    geom_boxplot(aes(fill = cut), color = 'navy') +
    theme_bw() +
    theme(legend.position = 'top')

Boxplot

Añadimos una 3ª variable, la variable ‘color’

ggplot(df, aes(x = cut, y = price)) +
    geom_boxplot(aes(fill = color), color = 'navy') +
    theme_bw() +
    theme(legend.position = 'top')

Boxplot

ggplot(df, aes(x = cut, y = price)) +
    geom_boxplot(aes(fill = color), color = 'navy') +
    theme_bw() +
    theme(legend.position = 'top') +
    facet_wrap(~ color)

Histogram

ggplot(df, aes(x = price)) +
    geom_histogram(color = 'navy', fill = 'steelblue') +
    theme_bw()

Histogram

ggplot(df, aes(x = price)) +
    geom_histogram(aes(color = color), fill = 'white') +
    theme_bw()

Bar Chart

ggplot(df, aes(x = clarity)) +
    geom_bar(color = 'navy', fill = 'white') +
    theme_bw()

Bar Chart

ggplot(df, aes(x = clarity)) +
    geom_bar(aes(color = clarity, fill = clarity)) +
    theme_bw() +
    theme(legend.position = 'top')

Bar Chart

ggplot(df, aes(x = clarity)) +
    geom_bar(aes(color = cut, fill = cut)) +
    theme_bw() +
    theme(legend.position = 'top')

Bar Chart

ggplot(df, aes(x = clarity)) +
    geom_bar(aes(color = cut, fill = cut), position = position_dodge()) +
    theme_bw() +
    theme(legend.position = 'top')

Bar Chart

ggplot(df, aes(x = clarity, y = price)) +
    geom_bar(color = 'navy', stat = 'identity') +
    theme_bw()

Jitter

ggplot(df, aes(x = cut, y = clarity)) +
    geom_jitter(aes(color = color)) +
    theme_bw()

Pie Chart

ggplot(df, aes(x = '', y = price)) +
    geom_bar(aes(fill = clarity), stat = 'identity') +
    theme_bw() +
    coord_polar('y')

Volcano Plot

Vamos a realizar el volcano plot correspondiente al análisis diferencial de los datos de Drosophila melanogaster. Eliminamos las filas con 0 o NA values

dge <- read.csv('data/DGE.csv')
dge <- dge[dge$padj != 0, ]
dge <- dge[complete.cases(dge), ]

head(dge)

            X baseMean log2FoldChange      lfcSE       pvalue         padj
1 FBgn0000008 562.4671    -0.05656922 0.05473453 3.005379e-01 3.499743e-01
2 FBgn0000014 894.8910    -0.79584159 0.04535355 3.664583e-69 2.568605e-68
3 FBgn0000015 323.3163    -0.84122112 0.07377132 2.177357e-30 8.697926e-30
4 FBgn0000017 862.1217    -0.25306214 0.04781241 1.098852e-07 2.259261e-07
5 FBgn0000018 111.7811     0.37605333 0.11157453 6.718636e-04 1.085105e-03
6 FBgn0000024 789.7766    -0.91597718 0.04844234 5.862028e-80 4.626392e-79

Volcano Plot

Seleccionamos los valores de FC = 2 y padj = 0.0000001 de corte.

padj.cutoff <- 0.0000001
lfc.cutoff <- log2(2)

dge$class <- 'none'
dge[dge$log2FoldChange >= lfc.cutoff & dge$padj <= padj.cutoff, c('class')] <- 'UP'
dge[dge$log2FoldChange <= -1 * lfc.cutoff & dge$padj <= padj.cutoff, c('class')] <- 'DOWN'

head(dge)

            X baseMean log2FoldChange      lfcSE       pvalue         padj
1 FBgn0000008 562.4671    -0.05656922 0.05473453 3.005379e-01 3.499743e-01
2 FBgn0000014 894.8910    -0.79584159 0.04535355 3.664583e-69 2.568605e-68
3 FBgn0000015 323.3163    -0.84122112 0.07377132 2.177357e-30 8.697926e-30
4 FBgn0000017 862.1217    -0.25306214 0.04781241 1.098852e-07 2.259261e-07
5 FBgn0000018 111.7811     0.37605333 0.11157453 6.718636e-04 1.085105e-03
6 FBgn0000024 789.7766    -0.91597718 0.04844234 5.862028e-80 4.626392e-79
  class
1  none
2  none
3  none
4  none
5  none
6  none

Volcano Plot

ggplot(dge, aes(x = log2FoldChange, y = -1 * log10(padj))) +
    geom_point() +
    theme_bw()

Volcano Plot

ggplot(dge, aes(x = log2FoldChange, y = -1 * log10(padj))) +
    geom_point() +
    geom_hline(yintercept = -1 * log10(padj.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    geom_vline(xintercept=c(-1 * lfc.cutoff ,lfc.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    theme_bw()

Volcano Plot

ggplot(dge, aes(x = log2FoldChange, y = -1 * log10(padj))) +
    geom_point(aes(color = class)) +
    geom_hline(yintercept = -1 * log10(padj.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    geom_vline(xintercept=c(-1 * lfc.cutoff ,lfc.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    theme_bw()

Volcano Plot

colors <- c("UP"="#FC4E07", "none"="#E7B800", "DOWN"="#00AFBB")
ggplot(dge, aes(x = log2FoldChange, y = -1 * log10(padj))) +
    geom_point(aes(color = class)) +
    scale_color_manual(values = colors) +
    geom_hline(yintercept = -1 * log10(padj.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    geom_vline(xintercept=c(-1 * lfc.cutoff ,lfc.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    theme_bw()

Volcano Plot

library(ggrepel)
ggplot(dge, aes(x = log2FoldChange, y = -1 * log10(padj), label = X)) +
    geom_point(aes(color = class)) +
    scale_color_manual(values = colors) +
    geom_hline(yintercept = -1 * log10(padj.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    geom_vline(xintercept=c(-1 * lfc.cutoff ,lfc.cutoff ), linetype="dashed", 
                              color = "black", linewidth = 0.2) +
    geom_label_repel(data = dge[dge$class %in% c('UP', 'DOWN'), ], 
                            size= 4, color = 'firebrick',  point.padding=unit(0.5, "lines"), max.overlaps = 30) +
    theme_bw()

Volcano Plot

TOP Genes

dge <- dge[order(dge$log2FoldChange, decreasing = TRUE), ]
genes.up <- dge[1:40, c('X')]
genes.down <- dge[(nrow(dge) - 30):nrow(dge), c('X')]

genes.up

 [1] "FBgn0039443" "FBgn0030921" "FBgn0035865" "FBgn0030570" "FBgn0040743"
 [6] "FBgn0039448" "FBgn0038394" "FBgn0032289" "FBgn0051626" "FBgn0051560"
[11] "FBgn0039027" "FBgn0039264" "FBgn0031467" "FBgn0036593" "FBgn0052855"
[16] "FBgn0003065" "FBgn0036596" "FBgn0033481" "FBgn0050471" "FBgn0031957"
[21] "FBgn0034131" "FBgn0038439" "FBgn0051561" "FBgn0051081" "FBgn0085359"
[26] "FBgn0051876" "FBgn0037940" "FBgn0036350" "FBgn0085250" "FBgn0035875"
[31] "FBgn0030841" "FBgn0035685" "FBgn0037395" "FBgn0032609" "FBgn0030107"
[36] "FBgn0039387" "FBgn0036532" "FBgn0004511" "FBgn0038002" "FBgn0029647"

genes.down

 [1] "FBgn0036417" "FBgn0032184" "FBgn0030830" "FBgn0005664" "FBgn0039476"
 [6] "FBgn0050334" "FBgn0085232" "FBgn0034092" "FBgn0037288" "FBgn0038007"
[11] "FBgn0038160" "FBgn0034828" "FBgn0031940" "FBgn0036470" "FBgn0039083"
[16] "FBgn0085363" "FBgn0039483" "FBgn0039299" "FBgn0033404" "FBgn0035022"
[21] "FBgn0019929" "FBgn0038148" "FBgn0038819" "FBgn0040553" "FBgn0035661"
[26] "FBgn0052453" "FBgn0038505" "FBgn0053270" "FBgn0031276" "FBgn0037179"
[31] "FBgn0037177"

TOP Genes

df.norm <- read.csv('data/normalized_counts.csv')
head(df.norm)

            X  SRX008026  SRX008174 SRX008201  SRX008239 SRX008008 SRX008168
1 FBgn0000008  577.30414  589.66391  536.2258  595.58471  560.6214  581.6257
2 FBgn0000014 1147.76523 1197.37875 1121.8777 1217.76428  671.0268  695.3146
3 FBgn0000015  395.03031  443.75218  452.2948  431.81641  229.7358  255.3881
4 FBgn0000017  895.81677 1007.84392  952.1506  930.81982  780.1100  792.5268
5 FBgn0000018   93.31425   96.27166   95.1218   97.28117  122.6359  154.8805
6 FBgn0000024 1117.90467 1063.50098 1050.0701 1048.39706  562.2742  561.8537
  SRX008211 SRX008255 SRX008261
1  516.9053  536.3239  567.9487
2  682.5334  668.7586  651.6000
3  236.6116  256.0892  209.1284
4  789.9186  771.1942  838.7149
5  120.1259  118.5327  107.8662
6  607.9097  578.7615  517.3176

df.norm <- df.norm[df.norm$X %in% c(genes.up, genes.down), ]
rownames(df.norm) <- df.norm$X
df.norm$X <- NULL
dim(df.norm)

[1] 71  9

Heatmap

library(pheatmap)

pheatmap(df.norm)

Heatmap

pheatmap(df.norm,
        scale = 'row'
)

Heatmap

pheatmap(df.norm,
        scale = 'row',
        cutree_cols =  2,
        cutree_rows = 2,
        color = colorRampPalette(
            c("steelblue", 
                "white", 
                "firebrick3")
            )(100),
)

Heatmap

pheatmap(df.norm,
        scale = 'row',
        cutree_cols =  2,
        cutree_rows = 2,
        color = colorRampPalette(
            c("steelblue", 
                "white", 
                "firebrick3")
            )(100),
        annotation_col = coldata
)

Volver