dyntoy simulates single-cell expression data in which a single-cell trajectory is present. Even though the model to generate the data is very simplistic (and far from realistic), it can simulate very complex trajectory models, such as large trees, convergences and loops.
As the data is relatively easy, it can be used to quickly test and prototype a TI method. However, for more realistic synthetic data, check out our dyngen package.
Install using devtools:
# install.packages("devtools")
devtools::install_github("dynverse/dyntoy")dyntoy contains some pre-generated toy data within the toy_datasets
data object:
data("toy_datasets", package = "dyntoy")Data can be generated using generate_dataset:
library(dyntoy)
dataset <- generate_dataset(
model = model_bifurcating(),
num_cells = 1000,
num_features = 1000
)
dataset$milestone_network
#> # A tibble: 3 x 4
#> from to length directed
#> <chr> <chr> <dbl> <lgl>
#> 1 M3 M4 0.191 TRUE
#> 2 M1 M3 0.417 TRUE
#> 3 M3 M2 0.0432 TRUE- PROSSTT: https://github.com/soedinglab/prosstt
- Splatter: https://github.com/Oshlack/splatter
Check out news(package = "dynwrap") or NEWS.md for a
full list of
changes.
-
Initial release of dyntoy
-
Generates simple toy datasets containg trajectories, with expression, RNA velocity, differentially expressed features
