| Title: | Create Muller Plots of Evolutionary Dynamics |
|---|---|
| Description: | Create plots that combine a phylogeny and frequency dynamics. Phylogenetic input can be a generic adjacency matrix or a tree of class "phylo". Inspired by similar plots in publications of the labs of RE Lenski and JE Barrick. Named for HJ Muller (who popularised such plots) and H Wickham (whose code this package exploits). |
| Authors: | Robert Noble [aut, cre] |
| Maintainer: | Robert Noble <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.6.0 |
| Built: | 2025-02-09 03:44:30 UTC |
| Source: | https://github.com/robjohnnoble/ggmuller |
The function adds rows at each time point recording the difference between the total population and its maximum value. Generally there is no need to use this function as Muller_pop_plot calls it automatically.
add_empty_pop(Muller_df)add_empty_pop(Muller_df)
Muller_df |
Dataframe created by get_Muller_df |
A dataframe that can be used as input in Muller_plot.
Rob Noble, [email protected]
Muller_df <- get_Muller_df(example_edges, example_pop_df) Muller_df2 <- add_empty_pop(Muller_df)Muller_df <- get_Muller_df(example_edges, example_pop_df) Muller_df2 <- add_empty_pop(Muller_df)
Add a row to the edges list to represent the root node (if not already present).
add_root_row(tree)add_root_row(tree)
tree |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
The same dataframe including a row representing the root node.
Rob Noble, [email protected]
tree1 <- data.frame(Parent = c(1,1,1,2,3,4), Identity = 2:7) add_root_row(tree1)tree1 <- data.frame(Parent = c(1,1,1,2,3,4), Identity = 2:7) add_root_row(tree1)
The function 1) identifies when genotypes first have non-zero populations; 2) copies all the rows of data for these time points; 3) modifies the copied rows by decreasing Generation and setting Population of the emerging genotypes to be close to zero; and then 4) adds the modified rows to the dataframe. This ensures that ggplot plots genotypes arising at the correct time points.
add_start_points(pop_df, start_positions = 0.5)add_start_points(pop_df, start_positions = 0.5)
pop_df |
Dataframe with column names "Identity", "Population", and either "Generation" or "Time" |
start_positions |
Numeric value between 0 and 1 that determines the times at which genotypes are assumed to have arisen (see examples) |
By default, the function assumes that each genotype arose half way between the latest time at which its population is zero and the earliest time at which its population is greater than zero. You can override this assumption using the start_positions parameter. If start_positions = 0 (respetively 1) then each genotype is assumed to have arisen at the earliest (respectively latest) time compatible with the data. Intermediate values are also permitted.
The input Dataframe with additional rows.
Rob Noble, [email protected]
pop1 <- data.frame(Generation = rep(1:5, each = 4), Identity = rep(1:4, 5), Population = c(1,0,0,0,1,1,0,0,1,1,1,0,1,1,1,1,1,1,1,1)) add_start_points(pop1) # to see the effect of changing start_positions, compare the Generation columns: add_start_points(pop1, 0) add_start_points(pop1, 1)pop1 <- data.frame(Generation = rep(1:5, each = 4), Identity = rep(1:4, 5), Population = c(1,0,0,0,1,1,0,0,1,1,1,0,1,1,1,1,1,1,1,1)) add_start_points(pop1) # to see the effect of changing start_positions, compare the Generation columns: add_start_points(pop1, 0) add_start_points(pop1, 1)
Create a tree object of class "phylo" from an adjacency matrix
adj_matrix_to_tree(edges)adj_matrix_to_tree(edges)
edges |
Dataframe comprising an adjacency matrix, in which the first column is the parent and the second is the daughter. |
A phylo object.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) tree <- adj_matrix_to_tree(edges1) class(tree)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) tree <- adj_matrix_to_tree(edges1) class(tree)
Single nodes are those with exactly one daughter. This function is required by adj_matrix_to_tree, since valid "phylo" objects cannot contain single nodes. If pre-existing branches lack lengths then these are set to 1.
branch_singles(edges)branch_singles(edges)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
A dataframe comprising the augmented adjacency matrix.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3), Identity = 2:5) branch_singles(edges1)edges1 <- data.frame(Parent = c(1,1,1,3), Identity = 2:5) branch_singles(edges1)
Example dataframe containing both phylogenetic information and population dynamics.
data(example_df)data(example_df)
A dataframe with column names "Generation", "Identity", "Parent", "Population" and "RelativeFitness"
Example dataframe comprising an adjacency matrix.
data(example_edges)data(example_edges)
A dataframe with column names "Parent" and "Identity"
Example dataframe containing population dynamics.
data(example_pop_df)data(example_pop_df)
A dataframe with column names "Generation", "Identity" and "Population"
Returns the Parent value of the common ancestor.
find_start_node(edges)find_start_node(edges)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
The Parent that is the common ancestor.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) find_start_node(edges1)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) find_start_node(edges1)
Get adjacency list of a tree.
get_Adj(tree)get_Adj(tree)
tree |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
The adjacency list.
Rob Noble, [email protected]
tree1 <- data.frame(Parent = c(1,1,1,1,2,3,4), Identity = 1:7, Population = c(1, rep(5, 6))) get_Adj(tree1)tree1 <- data.frame(Parent = c(1,1,1,1,2,3,4), Identity = 1:7, Population = c(1, rep(5, 6))) get_Adj(tree1)
Extract an adjacency matrix from a larger data frame
get_edges(df, generation = NA)get_edges(df, generation = NA)
df |
Dataframe inclduing column names "Identity", "Parent", and either "Generation" or "Time" |
generation |
Numeric value of Generation (or Time) at which to determine the adjacency matrix (defaults to final time point) |
A dataframe comprising the adjacency matrix.
Rob Noble, [email protected]
## Not run: edges <- get_edges(example_df) # extract the adjacency matrix from the data frame: pop_df <- get_population_df(example_df) # create data frame for plot: Muller_df <- get_Muller_df(edges, pop_df) require(RColorBrewer) # for the palette # draw plot: num_cols <- length(unique(Muller_df$RelativeFitness)) + 1 Muller_df$RelativeFitness <- as.factor(Muller_df$RelativeFitness) Muller_plot(Muller_df, colour_by = "RelativeFitness", palette = rev(colorRampPalette(brewer.pal(9, "YlOrRd"))(num_cols)), add_legend = TRUE) ## End(Not run)## Not run: edges <- get_edges(example_df) # extract the adjacency matrix from the data frame: pop_df <- get_population_df(example_df) # create data frame for plot: Muller_df <- get_Muller_df(edges, pop_df) require(RColorBrewer) # for the palette # draw plot: num_cols <- length(unique(Muller_df$RelativeFitness)) + 1 Muller_df$RelativeFitness <- as.factor(Muller_df$RelativeFitness) Muller_plot(Muller_df, colour_by = "RelativeFitness", palette = rev(colorRampPalette(brewer.pal(9, "YlOrRd"))(num_cols)), add_legend = TRUE) ## End(Not run)
Create a data frame from which to create a Muller plot
get_Muller_df( edges, pop_df, cutoff = 0, start_positions = 0.5, threshold = NA, add_zeroes = NA, smooth_start_points = NA )get_Muller_df( edges, pop_df, cutoff = 0, start_positions = 0.5, threshold = NA, add_zeroes = NA, smooth_start_points = NA )
edges |
Dataframe comprising an adjacency matrix, or tree of class "phylo" |
pop_df |
Dataframe with column names "Identity", "Population", and either "Generation" or "Time" |
cutoff |
Numeric cutoff; genotypes that never become more abundant than this value are omitted |
start_positions |
Numeric value between 0 and 1 that determines the times at which genotypes are assumed to have arisen (see examples) |
threshold |
Depcrecated (use cutoff instead, but note that "threshold" omitted genotypes that never become more abundant than *twice* its value) |
add_zeroes |
Deprecated (now always TRUE) |
smooth_start_points |
Deprecated (now always TRUE) |
A dataframe that can be used as input in Muller_plot and Muller_pop_plot.
Rob Noble, [email protected]
# by default, all genotypes are included, # but one can choose to omit genotypes with max frequency < cutoff: Muller_df <- get_Muller_df(example_edges, example_pop_df, cutoff = 0.01) # the genotype names can be arbitrary character strings instead of numbers: example_edges_char <- example_edges example_edges_char$Identity <- paste0("foo", example_edges_char$Identity, "bar") example_edges_char$Parent <- paste0("foo", example_edges_char$Parent, "bar") example_pop_df_char <- example_pop_df example_pop_df_char$Identity <- paste0("foo", example_pop_df_char$Identity, "bar") Muller_df <- get_Muller_df(example_edges_char, example_pop_df_char, cutoff = 0.01) # the genotype names can also be factors (which is the default for strings in imported data): example_edges_char$Identity <- as.factor(example_edges_char$Identity) example_edges_char$Parent <- as.factor(example_edges_char$Parent) example_pop_df_char$Identity <- as.factor(example_pop_df_char$Identity) Muller_df <- get_Muller_df(example_edges_char, example_pop_df_char, cutoff = 0.01) # to see the effect of changing start_positions, compare these two plots: edges1 <- data.frame(Parent = c(1,2,1), Identity = 2:4) pop1 <- data.frame(Time = rep(1:4, each = 4), Identity = rep(1:4, times = 4), Population = c(1, 0, 0, 0, 2, 2, 0, 0, 4, 8, 4, 0, 8, 32, 32, 16)) df0 <- get_Muller_df(edges1, pop1, start_positions = 0) df1 <- get_Muller_df(edges1, pop1, start_positions = 1) Muller_plot(df0) Muller_plot(df1)# by default, all genotypes are included, # but one can choose to omit genotypes with max frequency < cutoff: Muller_df <- get_Muller_df(example_edges, example_pop_df, cutoff = 0.01) # the genotype names can be arbitrary character strings instead of numbers: example_edges_char <- example_edges example_edges_char$Identity <- paste0("foo", example_edges_char$Identity, "bar") example_edges_char$Parent <- paste0("foo", example_edges_char$Parent, "bar") example_pop_df_char <- example_pop_df example_pop_df_char$Identity <- paste0("foo", example_pop_df_char$Identity, "bar") Muller_df <- get_Muller_df(example_edges_char, example_pop_df_char, cutoff = 0.01) # the genotype names can also be factors (which is the default for strings in imported data): example_edges_char$Identity <- as.factor(example_edges_char$Identity) example_edges_char$Parent <- as.factor(example_edges_char$Parent) example_pop_df_char$Identity <- as.factor(example_pop_df_char$Identity) Muller_df <- get_Muller_df(example_edges_char, example_pop_df_char, cutoff = 0.01) # to see the effect of changing start_positions, compare these two plots: edges1 <- data.frame(Parent = c(1,2,1), Identity = 2:4) pop1 <- data.frame(Time = rep(1:4, each = 4), Identity = rep(1:4, times = 4), Population = c(1, 0, 0, 0, 2, 2, 0, 0, 4, 8, 4, 0, 8, 32, 32, 16)) df0 <- get_Muller_df(edges1, pop1, start_positions = 0) df1 <- get_Muller_df(edges1, pop1, start_positions = 1) Muller_plot(df0) Muller_plot(df1)
Extract population data from a larger data frame
get_population_df(df)get_population_df(df)
df |
Dataframe inclduing column names "Identity", "Parent", and either "Generation" or "Time" |
A dataframe comprising the population dynamics.
Rob Noble, [email protected]
## Not run: # extract the adjacency matrix from the data frame: edges <- get_edges(example_df) # extract the populations (and any other attributes) from the data frame: pop_df <- get_population_df(example_df) # create data frame for plot: Muller_df <- get_Muller_df(edges, pop_df) require(RColorBrewer) # for the palette # draw plot: num_cols <- length(unique(Muller_df$RelativeFitness)) + 1 Muller_df$RelativeFitness <- as.factor(Muller_df$RelativeFitness) Muller_plot(Muller_df, colour_by = "RelativeFitness", palette = rev(colorRampPalette(brewer.pal(9, "YlOrRd"))(num_cols)), add_legend = TRUE) ## End(Not run)## Not run: # extract the adjacency matrix from the data frame: edges <- get_edges(example_df) # extract the populations (and any other attributes) from the data frame: pop_df <- get_population_df(example_df) # create data frame for plot: Muller_df <- get_Muller_df(edges, pop_df) require(RColorBrewer) # for the palette # draw plot: num_cols <- length(unique(Muller_df$RelativeFitness)) + 1 Muller_df$RelativeFitness <- as.factor(Muller_df$RelativeFitness) Muller_plot(Muller_df, colour_by = "RelativeFitness", palette = rev(colorRampPalette(brewer.pal(9, "YlOrRd"))(num_cols)), add_legend = TRUE) ## End(Not run)
Returns the first Identity value in the sorted set of daughters. When parent has no daughters, returns the input Identity.
move_down(edges, parent)move_down(edges, parent)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
parent |
number or character string specifying whose daughter is to be found |
The daughter's Identity.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_down(edges1, 3)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_down(edges1, 3)
Returns the next Identity value among the sorted set of siblings. When there is no such sibling, returns the input Identity.
move_right(edges, identity)move_right(edges, identity)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
identity |
number or character string specifying whose sibling is to be found |
The sibling's Identity.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_right(edges1, 3)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_right(edges1, 3)
Returns the corresponding Parent value. When there is no parent (i.e. at the top of the tree), returns the input Identity.
move_up(edges, identity)move_up(edges, identity)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
identity |
number or character string specifying daughter whose parent is to be found |
The Parent value.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_up(edges1, 3)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) move_up(edges1, 3)
Draw a Muller plot of frequencies using ggplot2
Muller_plot( Muller_df, colour_by = "Identity", palette = NA, add_legend = FALSE, xlab = NA, ylab = "Frequency", pop_plot = FALSE, conceal_edges = FALSE )Muller_plot( Muller_df, colour_by = "Identity", palette = NA, add_legend = FALSE, xlab = NA, ylab = "Frequency", pop_plot = FALSE, conceal_edges = FALSE )
Muller_df |
Dataframe created by get_Muller_df |
colour_by |
Character containing name of column by which to colour the plot |
palette |
Either a brewer palette or a vector of colours (if colour_by is categorical) |
add_legend |
Logical whether to show legend |
xlab |
Label of x axis |
ylab |
Label of y axis |
pop_plot |
Logical for whether this function is being called from Muller_pop_plot (otherwise should be FALSE) |
conceal_edges |
Whether try to conceal the edges between polygons (usually unnecessary or undesirable) |
None
Rob Noble, [email protected]
# include all genotypes: Muller_df1 <- get_Muller_df(example_edges, example_pop_df) Muller_plot(Muller_df1) # omit genotypes with max frequency < 0.1: Muller_df2 <- get_Muller_df(example_edges, example_pop_df, cutoff = 0.2) Muller_plot(Muller_df2) # colour by a continuous variable: Muller_df1 <- get_Muller_df(example_edges, example_pop_df) Muller_df1$Val <- as.numeric(Muller_df1$Identity) Muller_plot(Muller_df1, colour_by = "Val", add_legend = TRUE)# include all genotypes: Muller_df1 <- get_Muller_df(example_edges, example_pop_df) Muller_plot(Muller_df1) # omit genotypes with max frequency < 0.1: Muller_df2 <- get_Muller_df(example_edges, example_pop_df, cutoff = 0.2) Muller_plot(Muller_df2) # colour by a continuous variable: Muller_df1 <- get_Muller_df(example_edges, example_pop_df) Muller_df1$Val <- as.numeric(Muller_df1$Identity) Muller_plot(Muller_df1, colour_by = "Val", add_legend = TRUE)
This variation on the Muller plot, which shows variation in population size as well as frequency, is also known as a fish plot.
Muller_pop_plot( Muller_df, colour_by = "Identity", palette = NA, add_legend = FALSE, xlab = NA, ylab = "Population", conceal_edges = FALSE )Muller_pop_plot( Muller_df, colour_by = "Identity", palette = NA, add_legend = FALSE, xlab = NA, ylab = "Population", conceal_edges = FALSE )
Muller_df |
Dataframe created by get_Muller_df |
colour_by |
Character containing name of column by which to colour the plot |
palette |
Either a brewer palette or a vector of colours (if colour_by is categorical) |
add_legend |
Logical whether to show legend |
xlab |
Label of x axis |
ylab |
Label of y axis |
conceal_edges |
Whether try to conceal the edges between polygons (usually unnecessary or undesirable) |
None
Rob Noble, [email protected]
Muller_df <- get_Muller_df(example_edges, example_pop_df) Muller_pop_plot(Muller_df)Muller_df <- get_Muller_df(example_edges, example_pop_df) Muller_pop_plot(Muller_df)
Nodes are traversed in the order that they should be stacked in a Muller plot. Each node appears exactly twice.
path_vector(edges)path_vector(edges)
edges |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
A vector specifying the path.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) path_vector(edges1)edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) path_vector(edges1)
Nodes are traversed in the order that they should be stacked in a Muller plot. Each node appears exactly twice.
path_vector_new( tree, i = NULL, Adj = NULL, Col = NULL, is_leaf = NULL, path = NULL )path_vector_new( tree, i = NULL, Adj = NULL, Col = NULL, is_leaf = NULL, path = NULL )
tree |
Dataframe comprising an adjacency matrix, with column names "Parent" and "Identity" |
i |
Current node |
Adj |
Adjacency matrix |
Col |
Node label |
is_leaf |
Label whether node is a leaf |
path |
The path vector so far |
A list, including a vector specifying the path.
Rob Noble, [email protected]
edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) path_vector_new(edges1)$pathedges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) path_vector_new(edges1)$path
Reorder a Muller plot dataframe by a vector
reorder_by_vector(df, vector)reorder_by_vector(df, vector)
df |
Dataframe with column names "Identity", "Parent", and either "Generation" or "Time", in which each Identity appears exactly twice |
vector |
Vector of Identity values |
The reordered dataframe.
Rob Noble, [email protected]
df <- data.frame(Generation = c(rep(0, 6), rep(1, 6)), Identity = rep(1:6,2), Population = c(1, rep(0, 5), 10, rep(1, 5))) df <- rbind(df, df) # duplicate rows require(dplyr) df <- arrange(df, Generation) # put in chronological order edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) # adjacency matrix path <- path_vector_new(edges1)$path # path through the adjacency matrix reorder_by_vector(df, path)df <- data.frame(Generation = c(rep(0, 6), rep(1, 6)), Identity = rep(1:6,2), Population = c(1, rep(0, 5), 10, rep(1, 5))) df <- rbind(df, df) # duplicate rows require(dplyr) df <- arrange(df, Generation) # put in chronological order edges1 <- data.frame(Parent = c(1,1,1,3,3), Identity = 2:6) # adjacency matrix path <- path_vector_new(edges1)$path # path through the adjacency matrix reorder_by_vector(df, path)