Package: PathwaySpace 1.1.0

Highlights

Produces landscape images representing graphs by geodesic paths
Projects signals using a decay function to model signal attenuation
Applies a convolution algorithm to combine signals from neighboring vertices

Overview

For a given igraph object containing vertices, edges, and a signal associated with the vertices, PathwaySpace performs a convolution operation, which involves a weighted combination of neighboring signals on a graph. Figure 1A illustrates the convolution operation problem. Each vertex’s signal is positioned on a grid at specific x and y coordinates, represented by cones (for available signals) or question marks (for null or missing values).

Figure 1. Signal processing addressed by the PathwaySpace package. A) Graph overlaid on a 2D coordinate system. Each projection cone represents the signal associated with a graph vertex (referred to as vertex-signal positions), while question marks indicate positions with no signal information (referred to as null-signal positions). Inset: Graph layout of the toy example used in the quick start section of this vignette. B) Illustration of signal projection from two neighboring vertices, simplified to one dimension. Right: Signal profiles from aggregation and decay functions.

Our model considers the vertex-signal positions as source points (or transmitters) and the null-signal positions as end points (or receivers). The signal values from vertex-signal positions are then projected to the null-signal positions according to a decay function, which will control how the signal values attenuate as they propagate across the 2D space. For a given null-signal position, the k-top signals are used to define the contributing vertices for the convolution operation, which will aggregate the signals from these contributing vertices considering their intensities reaching the end points. Users can adjust both the aggregation and decay functions; the aggregation function can be any arithmetic rule that reduces a numeric vector into a single scalar value (e.g., mean, weighted mean), while available decay functions include linear, exponential, and Weibull models (Fig.1B). Additionally, users can assign vertex-specific decay functions to model signal projections for subsets of vertices that may exhibit distinct behaviors. The resulting image forms geodesic paths in which the signal has been projected from vertex- to null-signal positions, using a density metric to measure the signal intensity along these paths.

Setting basic input data

#--- Load required packages for this section
library(igraph)
library(ggplot2)
library(RGraphSpace)
library(PathwaySpace)

This section will create an igraph object containing a binary signal associated to each vertex. The graph layout is configured manually to ensure that users can easily view all the relevant arguments needed to prepare the input data for the PathwaySpace package. The igraph’s make_star() function creates a star-like graph and the V() function is used to set attributes for the vertices. The PathwaySpace package will require that all vertices have x, y, and name attributes.

# Make a 'toy' igraph object, either a directed or undirected graph
gtoy1 <- make_star(5, mode="undirected")

# Assign 'x' and 'y' coordinates to each vertex
# ..this can be an arbitrary unit in (-Inf, +Inf)
V(gtoy1)$x <- c(0, 2, -2, -4, -8)
V(gtoy1)$y <- c(0, 0,  2, -4,  0)

# Assign a 'name' to each vertex (here, from n1 to n5)
V(gtoy1)$name <- paste0("n", 1:5)

Checking graph validity

Next, we will create a GraphSpace-class object using the GraphSpace() constructor. This function will check the validity of the igraph object. For this example mar = 0.2, which sets the outer margins of the graph.

# Check graph validity
g_space1 <- GraphSpace(gtoy1, mar = 0.2)

Our graph is now ready for the PathwaySpace package. We can check its layout using the plotGraphSpace() function.

# Check the graph layout
plotGraphSpace(g_space1, add.labels = TRUE)

Creating a PathwaySpace object

Next, we will create a PathwaySpace-class object using the buildPathwaySpace() constructor. This will calculate pairwise distances between vertices, subsequently required by the signal projection methods.

# Run the PathwaySpace constructor
p_space1 <- buildPathwaySpace(g_space1)

As a default behavior, the buildPathwaySpace() constructor initializes the signal of each vertex as 0. We can use the vertexSignal() accessor to get and set vertex signals in a PathwaySpace object; for example, in order to get vertex names and signal values:

# Check the number of vertices in the PathwaySpace object
length(p_space1)
## [1] 5

# Check vertex names
names(p_space1)
## [1] "n1" "n2" "n3" "n4" "n5"

# Check signal (initialized with '0')
vertexSignal(p_space1)
## n1 n2 n3 n4 n5 
##  0  0  0  0  0

…and for setting new signal values in PathwaySpace objects:

# Set new signal to all vertices
vertexSignal(p_space1) <- c(1, 4, 2, 4, 3)

# Set a new signal to the 1st vertex
vertexSignal(p_space1)[1] <- 2

# Set a new signal to vertex "n1"
vertexSignal(p_space1)["n1"] <- 6

# Check updated signal values
vertexSignal(p_space1)
## n1 n2 n3 n4 n5 
##  6  4  2  4  3

Signal projection

Circular projection

Following that, we will use the circularProjection() function to project the network signals by the weibullDecay() function with pdist = 0.4, which is passed by the decay.fun argument. This term determines a distance unit for the signal convolution, affecting the extent over which the convolution operation projects the signal. For example, when pdist = 1, it will represent the diameter of the inscribed circle within the coordinate space. We also set k = 1, which defines the contributing vertices for signal convolution.

# Run signal projection
p_space1 <- circularProjection(p_space1, k = 1, 
  decay.fun = weibullDecay(pdist = 0.4))

# Plot a PathwaySpace image
plotPathwaySpace(p_space1, add.marks = TRUE)

Next, we reassess the same PathwaySpace object, using pdist = 0.2, k = 2 and adjusting the shape of the decay function (for further details, see the online tutorials).

# Re-run signal projection, adjusting Weibull's shape
p_space1 <- circularProjection(p_space1, k = 2, 
  decay.fun = weibullDecay(shape = 2, pdist = 0.2))

# Plot PathwaySpace
plotPathwaySpace(p_space1, marks = "n1", theme = "th2")

The shape parameter allows a projection to take a variety of shapes. When shape = 1 the projection follows an exponential decay, and when shape > 1 the projection is first convex, then concave with an inflection point along the decay path. For additional examples see the modeling signal decay tutorial.

Polar projection

In this section we will project network signals using a polar coordinate system. This representation may be useful for certain types of data, for example, to highlight patterns of signal propagation on directed graphs, especially to explore the orientation aspect of signal flow. To demonstrate this feature we will used the gtoy2 directed graph, available in the RGraphSpace package.

# Load a pre-processed directed igraph object
data("gtoy2", package = "RGraphSpace")
# Check graph validity
g_space2 <- GraphSpace(gtoy2, mar = 0.2)

# Check the graph layout
plotGraphSpace(g_space2, add.labels = TRUE)

# Build a PathwaySpace for the 'g_space2'
p_space2 <- buildPathwaySpace(g_space2)

# Set '1s' as vertex signal
vertexSignal(p_space2) <- 1

For fine-grained modeling of signal decay, the vertexDecay() accessor allows assigning decay functions at the level of individual vertices. For example, adjusting Weibull’s shape argument for node n6:

# Modify decay function
# ..for all vertices
vertexDecay(p_space2) <- weibullDecay(shape=2, pdist = 1)
# ..for individual vertices
vertexDecay(p_space2)[["n6"]] <- weibullDecay(shape=3, pdist = 1)

In polar projections, the pdist term defines a reference distance related to edge length, aiming to constrain signal projections within edge bounds. Here we set pdist = 1 to reach full edge lengths. Next, we run the signal projection using polar coordinates. The beta exponent will control the angular span; for values greater than zero, beta will progressively narrow the projection along the edge axis.

# Run signal projection using polar coordinates
p_space2 <- polarProjection(p_space2, beta = 10)

# Plot PathwaySpace
plotPathwaySpace(p_space2, theme = "th2", add.marks = TRUE)

Note that this projection distributes signals on the edges regardless of direction. To incorporate edge orientation, we set directional = TRUE, which channels the projection along the paths:

# Re-run signal projection using 'directional = TRUE'
p_space2 <- polarProjection(p_space2, beta = 10, directional = TRUE)

# Plot PathwaySpace
plotPathwaySpace(p_space2, theme = "th2", marks = c("n1","n3","n4","n5"))

This PathwaySpace polar projection emphasizes the signal flow along the directional pattern of a directed graph (see the igraph plot above). When interpreting, users should note that this approach introduces simplifications; for example, depending on the network topology, the polar projection may fail to capture complex features of directed graphs, such as cyclic dependencies, feedforward and feedback loops, or other intricate interactions.

Signal types

The PathwaySpace accepts binary, integer, and numeric signal types, including NAs. If a vertex signal is assigned with NA, it will be ignored by the convolution algorithm. Logical values are also allowed, but it will be treated as binary. Next, we show the projection of a signal that includes negative values, using the p_space1 object created previously.

# Set a negative signal to vertices "n3" and "n4"
vertexSignal(p_space1)[c("n3","n4")] <- c(-2, -4)

# Check updated signal vector
vertexSignal(p_space1)
# n1 n2 n3 n4 n5 
#  6  4 -2 -4  3 

# Re-run signal projection
p_space1 <- circularProjection(p_space1, decay.fun = weibullDecay(shape = 2))

# Plot PathwaySpace
plotPathwaySpace(p_space1, bg.color = "white", font.color = "grey20", add.marks = TRUE, mark.color = "magenta", theme = "th2")

Note that the original signal vector was rescale to [-1, +1]. If the signal vector is >=0, then it will be rescaled to [0, 1]; if the signal vector is <=0, it will be rescaled to [-1, 0]; and if the signal vector is in (-Inf, +Inf), then it will be rescaled to [-1, +1]. To override this signal processing, simply set rescale = FALSE in the projection function.

Online tutorials

https://sysbiolab.github.io/PathwaySpace/

Citation

If you use PathwaySpace, please cite:

Tercan & Apolonio et al. Protocol for assessing distances in pathway space for classifier feature sets from machine learning methods. STAR Protocols 6(2):103681, 2025. https://doi.org/10.1016/j.xpro.2025.103681
Ellrott et al. Classification of non-TCGA cancer samples to TCGA molecular subtypes using compact feature sets. Cancer Cell 43(2):195-212.e11, 2025. https://doi.org/10.1016/j.ccell.2024.12.002

Session information

## R version 4.5.2 (2025-10-31)
## Platform: x86_64-pc-linux-gnu
## Running under: Ubuntu 24.04.3 LTS
## 
## Matrix products: default
## BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
## LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so;  LAPACK version 3.12.0
## 
## locale:
##  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
##  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=C              
##  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
##  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
##  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
## 
## time zone: America/Sao_Paulo
## tzcode source: system (glibc)
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] PathwaySpace_1.1.0 RGraphSpace_1.1.0  ggplot2_4.0.0.9000 igraph_2.2.0      
## 
## loaded via a namespace (and not attached):
##  [1] vctrs_0.6.5        cli_3.6.5          knitr_1.50         rlang_1.1.6       
##  [5] xfun_0.53          ggrepel_0.9.6      generics_0.1.4     S7_0.2.0          
##  [9] jsonlite_2.0.0     glue_1.8.0         htmltools_0.5.8.1  sass_0.4.10       
## [13] scales_1.4.0       rmarkdown_2.30     grid_4.5.2         tibble_3.3.0      
## [17] evaluate_1.0.5     jquerylib_0.1.4    fastmap_1.2.0      yaml_2.3.10       
## [21] lifecycle_1.0.4    compiler_4.5.2     dplyr_1.1.4        RColorBrewer_1.1-3
## [25] Rcpp_1.1.0         pkgconfig_2.0.3    rstudioapi_0.17.1  farver_2.1.2      
## [29] digest_0.6.37      R6_2.6.1           RANN_2.6.2         tidyselect_1.2.1  
## [33] pillar_1.11.1      magrittr_2.0.4     bslib_0.9.0        withr_3.0.2       
## [37] tools_4.5.2        gtable_0.3.6       cachem_1.1.0

Getting started with PathwaySpace projection methods

Sysbiolab Team

2025-11-05