Chapter 11 Visualising biomolecular data

11.1 Heatmaps

Copy link

library("pRolocdata")
data(mulvey2015norm_se)
heatmap(assay(mulvey2015norm_se))

Figure 11.1: Heatmap of the (normalised) Mulvey et al. 2015 proteomic data.

A heatmap is composed of two hierarchical clusters (one along the rows, one along the columns, leading to their re-ordering based on their similarity) and a intensity matrix. Each of these components is subject to parameters and options.

As we have seen above, the distance used for clustering can have a substantial effect on the results, which is conformed below.

heatmap(assay(mulvey2015norm_se), distfun = cor.dist)

Figure 11.2: Heatmap of the (normalised) Mulvey et al. 2015 proteomic data using correlation distances.

Another important argument, scale controls whether rows, columns or none are scaled. Let’s re-use the toy data from the hierarchical clustering section below.

library("rWSBIM1322")
data(g3)
heatmap(g3, scale = "none", main = "No scaling")

Figure 11.3: Effect of scaling on the heatmap output: no scaling.

library("rWSBIM1322")
data(g3)
heatmap(g3, scale = "row", main = "Scaling along rows")

Figure 11.4: Effect of scaling on the heatmap output: scaling along the rows.

library("rWSBIM1322")
data(g3)
heatmap(g3, scale = "column", main = "Scaling along columns")

Figure 11.5: Effect of scaling on the heatmap output: scaling along the columns.

Based on the caveats above, it is essential to present and interpret heatmaps with great care.

There exists several packages that allow to produce heatmaps with various levels of sophistication, such as heatmap.2 from the r CRANpkg("gplots") package, the Heatplus package, or the ComplexHeatmap packages (full documentation here), demonstrated below.

library("ComplexHeatmap")
x <- assay(mulvey2015norm_se)
hcl <- hclust(dist(x))
cl <- cutree(hcl, k = 12)
ha1 <- HeatmapAnnotation(time = mulvey2015norm_se$times)
ha2 <- HeatmapAnnotation(boxplot = anno_boxplot(x))
ha3 <- rowAnnotation(cluster = factor(cl))
Heatmap(x,
        top_annotation = ha1,
        bottom_annotation = ha2,
        column_names_gp = gpar(fontsize = 8),
        row_names_gp = gpar(fontsize = 3)) +
    ha3

Figure 11.6: An annotated heatmap produced with the ComplexHeatmap Bioconductor package.

Other powerful packages to generate and customise heatmaps are superheat and pheatmap.

Finally, the heatmaply, d3heatmap and iheatmapr packages can be used to generate interactive heatmaps.

## prepare the data
library("pRolocdata")
data(hyperLOPIT2015_se)
## interactive heatmap
library("heatmaply")
heatmaply(assay(hyperLOPIT2015_se)[1:100, ])

heatmaply(assay(hyperLOPIT2015_se)[1:100, ],
          RowSideColors = as.numeric(as.factor(rowData(hyperLOPIT2015_se)$markers[1:100])))

library("d3heatmap")
d3heatmap(assay(hyperLOPIT2015_se))

See also A tutorial in displaying mass spectrometry-based proteomic data using heat maps (⊕Key 2012Key, M. 2012. “A Tutorial in Displaying Mass Spectrometry-Based Proteomic Data Using Heat Maps.” BMC Bioinformatics 13 Suppl 16: S10. https://doi.org/10.1186/1471-2105-13-S16-S10.), that applies to any type of omics data (not only proteomics) for a useful reference.

11.2 Visualising intersections between sets

Copy link

Computing and visualising intersections is a common task in data analysis. Venn and Euler diagrams are popular representation when comparing sets and their intersection. Two useful R packages to generate such plots are venneuler and Vennerable1515 You can install Vennerable with BiocManager::install("js229/Vennerable")..

We will use the mulvey2015norm_se feature names to generate a test data:

set.seed(123)
feat_list <- replicate(3,
                       sample(rownames(mulvey2015norm_se), 555),
                       simplify = FALSE)
names(feat_list) <- LETTERS[1:3]

The Venn function from the Vennerable package takes a list as input, and computes all possible intersections between these elements of the list. In the output below

• 000 refers to the empty set that are present in none of the element of the list;
• 001 is the set of items that are unique to the third element (named C) of our list;
• …
• 011 is the set of items that is shared by the second (B) and third (C) element (and absent from the first one) of our list;
• 111 is the set of items that are shared between all elements of our list.

library("Vennerable")
(v <- Venn(feat_list))

## A Venn object on 3 sets named
## A,B,C 
## 000 100 010 110 001 101 011 111 
##   0 336 317 105 340  82 101  32

Each of these intersections can be accessed using through the IntersectionSets slot.

## access the identifiers
v@IntersectionSets[["110"]]

##   [1] "Q8BXZ1"   "Q571H0"   "Q9D1C9"   "P43274"   "Q0VG62"   "Q8CE90-4"
##   [7] "Q99LI8"   "P30416"   "Q8C7X2-2" "Q9JI13-2" "Q8BJF9"   "Q6P5E4"  
##  [13] "P51881"   "Q8BGT7"   "Q8K2F0"   "P25206"   "Q921E6-3" "Q8BWR8"  
##  [19] "Q76MZ3"   "Q9WV32"   "O70493"   "Q62393-2" "Q9JIH2"   "Q9DCT2"  
##  [25] "Q80U49"   "P68254-2" "Q01730"   "Q9DBE9"   "Q9Z2U0"   "Q91WM2"  
##  [31] "Q8BNU0"   "Q8BRN9"   "Q08024-2" "Q8R1Q8"   "Q921F4"   "P60898"  
##  [37] "Q64213-2" "P36552"   "P61358"   "P83887"   "P61957"   "Q05816"  
##  [43] "Q9QXA5"   "Q52KI8-2" "Q62422"   "Q3UMF0-4" "Q9EQQ9"   "Q1PSW8"  
##  [49] "Q61183-4" "P08752"   "Q64511"   "P83940"   "Q60854"   "Q91W97"  
##  [55] "P24788"   "Q6ZWN5"   "Q9CQ62"   "Q9D7H3"   "Q8BGW1-3" "O88508"  
##  [61] "Q6PAR5-5" "P10107"   "Q99L04"   "Q8JZQ9"   "Q922S8"   "Q8C878"  
##  [67] "Q62241"   "Q9CZ15"   "P57780"   "P62897"   "Q9CY62"   "Q8R1N0"  
##  [73] "P63085"   "Q569Z5"   "Q9D0E1-2" "Q8BMP6"   "Q8CJ26"   "Q9R0P5"  
##  [79] "Q64373-2" "Q61464-4" "Q7TQK4"   "P61982"   "P51150"   "P84091"  
##  [85] "Q80XI4"   "Q8CFV9"   "Q8BTZ4-2" "Q8R326-2" "Q811J3"   "Q9D753"  
##  [91] "P51410"   "Q9D1H7"   "Q9CXK8"   "Q99LB6-2" "P62862"   "Q9WVB0"  
##  [97] "P62267"   "O88322"   "Q923E4"   "O54984"  
##  [ reached getOption("max.print") -- omitted 5 entries ]

And finally, the Venn object can directly be plotted (albeit with a suspicious set of colours) with

plot(v)

Figure 11.7: A Venn diagram representing the size of all intersections of the three elements of out feat_list input.

Venn diagrams are however limited to two to three, possibly four sets. The UpSetR package is a great solution when more sets need to be compared. The UpSetR visualises intersections of sets as a matrix in which the rows represent the sets and the columns represent their intersection sizes. For each set that is part of a given intersection, a black filled circle is placed in the corresponding matrix cell. If a set is not part of the intersection, a light grey circle is shown. A vertical black line connects the topmost black circle with the bottom most black circle in each column to emphasise the column-based relationships. The size of the intersections is shown as a bar chart placed on top of the matrix so that each column lines up with exactly one bar. A second bar chart showing the size of the each set is shown to the left of the matrix.

We will first make use of the fromList function to convert our list to a UpSetR compatible input and then generate the figure:

library("UpSetR")
upset_in <- fromList(feat_list)
upset(upset_in)

upset(upset_in, order.by = "freq")

upset(upset_in, order.by = "degree")

upset(upset_in, order.by = c("freq", "degree"))

upset(upset_in, order.by = c("degree", "freq"))

upset(upset_in, sets = c("A", "B"))

upset(upset_in, sets = c("B", "C", "A"), keep.order = TRUE)

upset(upset_in, group.by = "sets")

## Add set D with a single intersection
upset_in_4 <- upset_in
upset_in_4$D <- 0
upset_in_4[1, "D"] <- 1
head(upset_in_4)

##   A B C D
## 1 1 0 1 1
## 2 1 0 0 0
## 3 1 1 1 0
## 4 1 1 0 0
## 5 1 0 0 0
## 6 1 1 0 0

upset(upset_in_4)

upset(upset_in_4, empty.intersections = "on")

Visualising intersections with UpSetR shines with more than 4 sets, as Venn diagrams become practically useless.

There is also an UpSetR online app: https://gehlenborglab.shinyapps.io/upsetr/

set.seed(123)
feat_list_10 <- replicate(10, sample(rownames(mulvey2015norm_se), 555), simplify = FALSE)
names(feat_list_10) <- LETTERS[1:10]
feat_list_10 <- fromList(feat_list_10)
upset(feat_list_10)