,
Martin Pirkl [aut]| Title: | Pathway Enrichment Based on Differential Causal Effects |
|---|---|
| Description: | Compute differential causal effects (dce) on (biological) networks. Given observational samples from a control experiment and non-control (e.g., cancer) for two genes A and B, we can compute differential causal effects with a (generalized) linear regression. If the causal effect of gene A on gene B in the control samples is different from the causal effect in the non-control samples the dce will differ from zero. We regularize the dce computation by the inclusion of prior network information from pathway databases such as KEGG. |
| Authors: | Kim Philipp Jablonski [aut, cre]
,
Martin Pirkl [aut] |
| Maintainer: | Kim Philipp Jablonski <[email protected]> |
| License: | GPL-3 |
| Version: | 1.15.0 |
| Built: | 2024-10-31 01:07:47 UTC |
| Source: | https://github.com/bioc/dce |
From graphNEL with 0 edge weights to proper adjacency matrix
as_adjmat(g)as_adjmat(g)
g |
graphNEL object |
graph as adjacency matrix
dag <- create_random_DAG(30, 0.2) adj <- as_adjmat(dag)dag <- create_random_DAG(30, 0.2) adj <- as_adjmat(dag)
Turn dce object into data frame
## S3 method for class 'dce' as.data.frame(x, row.names = NULL, optional = FALSE, ...)## S3 method for class 'dce' as.data.frame(x, row.names = NULL, optional = FALSE, ...)
x |
dce object |
row.names |
optional character vector of rownames |
optional |
logical; allow optional arguments |
... |
additional arguments |
data frame containing the dce output
dag <- create_random_DAG(30, 0.2) X_wt <- simulate_data(dag) dag_mt <- resample_edge_weights(dag) X_mt <- simulate_data(dag_mt) dce_list <- dce(dag, X_wt, X_mt)dag <- create_random_DAG(30, 0.2) X_wt <- simulate_data(dag) dag_mt <- resample_edge_weights(dag) X_mt <- simulate_data(dag_mt) dce_list <- dce(dag, X_wt, X_mt)
Creates a DAG according to given parameters.
create_random_DAG( node_num, prob, eff_min = -1, eff_max = 1, node_labels = paste0("n", as.character(seq_len(node_num))), max_par = 3 )create_random_DAG( node_num, prob, eff_min = -1, eff_max = 1, node_labels = paste0("n", as.character(seq_len(node_num))), max_par = 3 )
node_num |
Number of nodes |
prob |
Probability of creating an edge |
eff_min |
Lower bound for edge weights |
eff_max |
Upper bound for edge weights |
node_labels |
Node labels |
max_par |
Maximal number of parents |
graph
Martin Pirkl
dag <- create_random_DAG(30, 0.2)dag <- create_random_DAG(30, 0.2)
Main function to compute differential causal effects and the pathway enrichment
dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'igraph' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'graphNEL' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'matrix' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO )dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'igraph' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'graphNEL' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO ) ## S4 method for signature 'matrix' dce( graph, df_expr_wt, df_expr_mt, solver = "lm", solver_args = list(), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO )
graph |
valid object defining a directed acyclic graph |
df_expr_wt |
data frame with wild type expression values |
df_expr_mt |
data from with mutation type expression values |
solver |
character with name of solver function |
solver_args |
additional arguments for the solver function. please adress this argument, if you use your own solver function. the default argument works with glm functions in the packages MASS, stats and glm2 |
adjustment_type |
character string for the method to define the adjustment set Z for the regression |
effect_type |
method of computing causal effects |
p_method |
character string. "mean", "sum" for standard summary functions, "hmp" for harmonic mean or any method from package 'metap', e.g., "meanp" or "sump". |
test |
either "wald" for testing significance with the wald test or "lr" for using a likelihood ratio test. Alternatively, "vcovHC" can improve results for zero-inflated date, i.e., from single cell RNAseq experiments. |
lib_size |
either a numeric vector of the same length as the sum of wild type and mutant samples or a logical. If TRUE, it is recommended that both data sets include not only the genes included in the graph but all genes available in the original data set. |
deconfounding |
indicates whether adjustment against latent confounding is used. If FALSE, no adjustment is used, if TRUE it adjusts for confounding by automatically estimating the number of latent confounders. The estimated number of latent confounders can be chosen manually by setting this variable to some number. |
conservative |
logical; if TRUE, does not use the indicator variable for the variables in the adjustment set |
log_level |
Control verbosity (logger::INFO, logger::DEBUG, ...) |
list of matrices with dces and corresponding p-value
dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce(dag,X.wt,X.mt)dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce(dag,X.wt,X.mt)
Shortcut for the main function to analyse negative binomial data
dce_nb( graph, df_expr_wt, df_expr_mt, solver_args = list(method = "glm.dce.nb.fit", link = "identity"), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO )dce_nb( graph, df_expr_wt, df_expr_mt, solver_args = list(method = "glm.dce.nb.fit", link = "identity"), adjustment_type = "parents", effect_type = "total", p_method = "hmp", test = "wald", lib_size = FALSE, deconfounding = FALSE, conservative = FALSE, log_level = logger::INFO )
graph |
valid object defining a directed acyclic graph |
df_expr_wt |
data frame with wild type expression values |
df_expr_mt |
data from with mutation type expression values |
solver_args |
additional arguments for the solver function |
adjustment_type |
character string for the method to define the adjustment set Z for the regression |
effect_type |
method of computing causal effects |
p_method |
character string. "mean", "sum" for standard summary functions, "hmp" for harmonic mean or any method from package 'metap', e.g., "meanp" or "sump". |
test |
either "wald" for testing significance with the wald test or "lr" for using a likelihood ratio test |
lib_size |
either a numeric vector of the same length as the sum of wild type and mutant samples or a logical. If TRUE, it is recommended that both data sets include not only the genes included in the graph but all genes available in the original data set. |
deconfounding |
indicates whether adjustment against latent confounding is used. If FALSE, no adjustment is used, if TRUE it adjusts for confounding by automatically estimating the number of latent confounders. The estimated number of latent confounders can be chosen manually by setting this variable to some number. |
conservative |
logical; if TRUE, does not use the indicator variable for the variables in the adjustment set |
log_level |
Control verbosity (logger::INFO, logger::DEBUG, ...) |
list of matrices with dces and corresponding p-value
dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce_nb(dag,X.wt,X.mt)dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce_nb(dag,X.wt,X.mt)
A dataset containing pathway statistics.
df_pathway_statisticsdf_pathway_statistics
A data frame with pathway statistics
Pathway database
Internal ID of pathway
Canonical name of pathway
Number of nodes in pathway
Number of edges in pathway
This function takes a DAG with edgeweights as input and computes the causal effects of all nodes on all direct and indirect children in the DAG. Alternatively see pcalg::causalEffect for pairwise computation.
estimate_latent_count(X1, X2, method = "auto")estimate_latent_count(X1, X2, method = "auto")
X1 |
data matrix corresponding to the first condition |
X2 |
data matrix corresponding to the second condition |
method |
a string indicating the method used for estimating the number of latent variables |
estimated number of latent variables
Domagoj Ćevid
graph1 <- create_random_DAG(node_num = 100, prob = .1) graph2 <- resample_edge_weights(graph1, tp=0.15) X1 <- simulate_data(graph1, n=200, latent = 3) X2 <- simulate_data(graph2, n=200, latent = 3) estimate_latent_count(X1, X2)graph1 <- create_random_DAG(node_num = 100, prob = .1) graph2 <- resample_edge_weights(graph1, tp=0.15) X1 <- simulate_data(graph1, n=200, latent = 3) X2 <- simulate_data(graph2, n=200, latent = 3) estimate_latent_count(X1, X2)
Converts a general graph to a dag with minimum distance to the original graph. The general idea is to transitively close the graph to detect cycles and remove them based on the rule "the more outgoing edges a node has, the more likely it is that incoming edges from a cycle will be deleted, and vice versa. However, this is too rigorous and deletes too many edges, which do not lead to a cycle. These edges are added back in the final step.
g2dag(g, tc = FALSE)g2dag(g, tc = FALSE)
g |
graph as adjacency matrix |
tc |
if TRUE computes the transitive closure |
dag as adjacency matrix
Ken Adams
g <- matrix(c(1,0,1,0, 1,1,0,0, 0,1,1,0, 1,1,0,1), 4, 4) rownames(g) <- colnames(g) <- LETTERS[seq_len(4)] dag <- g2dag(g)g <- matrix(c(1,0,1,0, 1,1,0,0, 0,1,1,0, 1,1,0,1), 4, 4) rownames(g) <- colnames(g) <- LETTERS[seq_len(4)] dag <- g2dag(g)
Dataframe containing meta-information of pathways in database
get_pathway_info( query_species = "hsapiens", database_list = NULL, include_network_statistics = FALSE )get_pathway_info( query_species = "hsapiens", database_list = NULL, include_network_statistics = FALSE )
query_species |
For which species |
database_list |
Which databases to query. Query all if 'NULL'. |
include_network_statistics |
Compute some useful statistics per pathway. Takes longer! |
data frame with pathway meta information
head(get_pathway_info(database_list = c("kegg")))head(get_pathway_info(database_list = c("kegg")))
Easy pathway network access
get_pathways( query_species = "hsapiens", database_list = NULL, remove_empty_pathways = TRUE, pathway_list = NULL )get_pathways( query_species = "hsapiens", database_list = NULL, remove_empty_pathways = TRUE, pathway_list = NULL )
query_species |
For which species |
database_list |
Which databases to query. Query all if 'NULL'. |
remove_empty_pathways |
Discard pathways without nodes |
pathway_list |
List mapping database name to vector of pathway names to download |
list of pathways
pathways <- get_pathways( pathway_list = list(kegg = c( "Protein processing in endoplasmic reticulum" )) ) plot_network(as(pathways[[1]]$graph, "matrix"))pathways <- get_pathways( pathway_list = list(kegg = c( "Protein processing in endoplasmic reticulum" )) ) plot_network(as(pathways[[1]]$graph, "matrix"))
Useful for performance evaluations
get_prediction_counts(truth, inferred, cutoff = 0.5)get_prediction_counts(truth, inferred, cutoff = 0.5)
truth |
Ground truth |
inferred |
Computed results |
cutoff |
Threshold for classification |
data.frame
Hans Wurst
get_prediction_counts(c(1,0), c(1,1))get_prediction_counts(c(1,0), c(1,1))
Create union of multiple graphs
graph_union(graph_list)graph_union(graph_list)
graph_list |
List of graphs |
graph union
dag <- create_random_DAG(30, 0.2) dag2 <- create_random_DAG(30, 0.2) graph_union(list(g1=dag,g2=dag2))dag <- create_random_DAG(30, 0.2) dag2 <- create_random_DAG(30, 0.2) graph_union(list(g1=dag,g2=dag2))
Convert graph object to dataframe with source and target columns
graph2df(graph)graph2df(graph)
graph |
Network |
data frame
dag <- create_random_DAG(30, 0.2) graph2df(dag)dag <- create_random_DAG(30, 0.2) graph2df(dag)
Robust partial correlation of column variables of a numeric matrix
pcor(x, g = NULL, adjustment_type = "parents", ...)pcor(x, g = NULL, adjustment_type = "parents", ...)
x |
matrix |
g |
related graph as adjacency matrix (optional) |
adjustment_type |
character string for the method to define the adjustment set Z for the regression |
... |
additional arguments for function 'cor' |
matrix of partial correlations
x <- matrix(rnorm(100),10,10) pcor(x)x <- matrix(rnorm(100),10,10) pcor(x)
Computes the significance of (partial) correlation based on permutations of the observations
permutation_test(x, y, iter = 1000, fun = pcor, mode = 1, ...)permutation_test(x, y, iter = 1000, fun = pcor, mode = 1, ...)
x |
wild type data set |
y |
mutant data set |
iter |
number of iterations (permutations) |
fun |
function to compute the statistic, e.g., cor or pcor |
mode |
either 1 for a function that takes a single data set and produces an output of class matrix, and 2, if the function takes two data sets |
... |
additional arguments for function 'fun' |
matrix of p-values
x <- matrix(rnorm(100),10,10) y <- matrix(rnorm(100),10,10) permutation_test(x,y,iter=10)x <- matrix(rnorm(100),10,10) y <- matrix(rnorm(100),10,10) permutation_test(x,y,iter=10)
Generic function which plots any adjacency matrix (assumes DAG)
plot_network( adja_matrix, nodename_map = NULL, edgescale_limits = NULL, nodesize = 17, labelsize = 3, node_color = "white", node_border_size = 0.5, arrow_size = 0.05, scale_edge_width_max = 1, show_edge_labels = FALSE, visualize_edge_weights = TRUE, use_symlog = FALSE, highlighted_nodes = c(), legend_title = "edge weight", value_matrix = NULL, shadowtext = FALSE, ... )plot_network( adja_matrix, nodename_map = NULL, edgescale_limits = NULL, nodesize = 17, labelsize = 3, node_color = "white", node_border_size = 0.5, arrow_size = 0.05, scale_edge_width_max = 1, show_edge_labels = FALSE, visualize_edge_weights = TRUE, use_symlog = FALSE, highlighted_nodes = c(), legend_title = "edge weight", value_matrix = NULL, shadowtext = FALSE, ... )
adja_matrix |
Adjacency matrix of network |
nodename_map |
node names |
edgescale_limits |
Limits for scale_edge_color_gradient2 (should contain 0). Useful to make plot comparable to others |
nodesize |
Node sizes |
labelsize |
Node label sizes |
node_color |
Which color to plot nodes in |
node_border_size |
Thickness of node's border stroke |
arrow_size |
Size of edge arrows |
scale_edge_width_max |
Max range for 'scale_edge_width' |
show_edge_labels |
Whether to show edge labels (DCEs) |
visualize_edge_weights |
Whether to change edge color/width/alpha relative to edge weight |
use_symlog |
Scale edge colors using dce::symlog |
highlighted_nodes |
List of nodes to highlight |
legend_title |
Title of edge weight legend |
value_matrix |
Optional matrix of edge weights if different from adjacency matrix |
shadowtext |
Draw white outline around node labels |
... |
additional parameters |
plot of dag and dces
Martin Pirkl, Kim Philipp Jablonski
adj <- matrix(c(0,0,0,1,0,0,0,1,0),3,3) plot_network(adj)adj <- matrix(c(0,0,0,1,0,0,0,1,0),3,3) plot_network(adj)
This function takes a differential causal effects object and plots the dag with the dces
## S3 method for class 'dce' plot(x, ...)## S3 method for class 'dce' plot(x, ...)
x |
dce object |
... |
Parameters passed to dce::plot_network |
plot of dag and dces
Martin Pirkl, Kim Philipp Jablonski
dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce.list <- dce(dag,X.wt,X.mt) plot(dce.list)dag <- create_random_DAG(30, 0.2) X.wt <- simulate_data(dag) dag.mt <- resample_edge_weights(dag) X.mt <- simulate_data(dag) dce.list <- dce(dag,X.wt,X.mt) plot(dce.list)
Remove non-gene nodes from pathway and reconnect nodes
propagate_gene_edges(graph)propagate_gene_edges(graph)
graph |
Biological pathway |
graph with only genes as nodes
dag <- create_random_DAG(30, 0.2) propagate_gene_edges(dag)dag <- create_random_DAG(30, 0.2) propagate_gene_edges(dag)
Takes a graph and modifies edge weights.
resample_edge_weights(g, tp = 0.5, mineff = 1, maxeff = 2, method = "unif")resample_edge_weights(g, tp = 0.5, mineff = 1, maxeff = 2, method = "unif")
g |
original graph |
tp |
fraction of edge weights which will be modified |
mineff |
minimal differential effect size |
maxeff |
maximum effect effect size or standard deviation, if method is "gauss" |
method |
method for drawing the differential for the causal effects. Can be "unif", "exp" or "gauss". |
graph with new edge weights
Martin Pirkl
graph.wt <- as(matrix(c(0,0,0,1,0,0,0,1,0), 3), "graphNEL") graph.mt <- resample_edge_weights(graph.wt)graph.wt <- as(matrix(c(0,0,0,1,0,0,0,1,0), 3), "graphNEL") graph.mt <- resample_edge_weights(graph.wt)
costum rlm function
rlm_dce(...)rlm_dce(...)
... |
see ?MASS::rlm |
Generate data for given DAG. The flexible framework allows for different distributions for source and child nodes. Default distributions are negative binomial (with mean = 100 and 1/dispersion = 100), and poisson, respectively.
simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'igraph' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'graphNEL' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'matrix' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" )simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'igraph' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'graphNEL' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" ) ## S4 method for signature 'matrix' simulate_data( graph, n = 100, dist_fun = rnbinom, dist_args = list(mu = 1000, size = 100), child_fun = rpois, child_args = list(), child_dep = "lambda", link_fun = negative.binomial.special()$linkfun, link_args = list(offset = 1), pop_size = 0, latent = 0, latent_fun = "unif" )
graph |
Graph to simulate on |
n |
Number of samples |
dist_fun |
distribution function for nodes without parents |
dist_args |
list of arguments for dist_fun |
child_fun |
distribution function for nodes with parents |
child_args |
list of arguments for child_fun |
child_dep |
link_fun computes an output for the expression of nodes without parents. this output is than used as input for child_fun. child_dep defines the parameter (a a string) of child_fun, which is used for the input. E.g., the link_fun is the identity and the child_fun is rnorm, we usually set child_dep = "mean". |
link_fun |
special link function for the negative binomial distribution |
link_args |
list of arguments for link_fun |
pop_size |
numeric for the population size, e.g., pop_size=1000 adds 1000-n random genes not in the graph |
latent |
number of latent variables |
latent_fun |
uniform "unif" or exponential "exp" distribution of latent coefficients |
graph
dag <- create_random_DAG(30, 0.2) X <- simulate_data(dag)dag <- create_random_DAG(30, 0.2) X <- simulate_data(dag)
summary for rlm_dce
## S3 method for class 'rlm_dce' summary(object, ...)## S3 method for class 'rlm_dce' summary(object, ...)
object |
object of class 'rlm_dce' |
... |
see ?MASS::summary.rlm |
Order rows/columns of a adjacency matrix topologically
topologically_ordering(adja_mat, alt = FALSE)topologically_ordering(adja_mat, alt = FALSE)
adja_mat |
Adjacency matrix of network |
alt |
Use igraph implementation |
topologically ordered matrix
adj <- matrix(c(0,1,0,0,0,1,0,0,0),3,3) topologically_ordering(adj)adj <- matrix(c(0,1,0,0,0,1,0,0,0),3,3) topologically_ordering(adj)
This function takes a DAG with edgeweights as input and computes the causal effects of all nodes on all direct and indirect children in the DAG. Alternatively see pcalg::causalEffect for pairwise computation.
trueEffects(g, partial = FALSE)trueEffects(g, partial = FALSE)
g |
graphNEL object |
partial |
if FALSE computes the total causal effects and not just the partial edge effects |
matrix of causal effects
Martin Pirkl
graph.wt <- as(matrix(c(0,0,0,1,0,0,0,1,0), 3), "graphNEL") trueEffects(graph.wt)graph.wt <- as(matrix(c(0,0,0,1,0,0,0,1,0), 3), "graphNEL") trueEffects(graph.wt)