ti_scfates_tree

A docker container for scFates, primarily for cell trajectory inference analysis of single-cell RNA-seq data.

The container was edited to follow the main scFates process, retaining some of the parameter settings, especially the tree learning methods, including 'Elastic Principal Graph (epg)' and 'simpleppt (ppt)'. Notably, we did not keep cellrank-related functions in it.

scFates repository and help document

https://github.com/LouisFaure/scFates

https://scfates.readthedocs.io/en/latest/

pull container

docker pull renjun0324/ti_scfates_tree

quick start

library(dynwrap)
library(dynmethods)
library(dyntoy)
library(tidyverse)
library(purrr)
library(dyno)

data("fibroblast_reprogramming_treutlein")

dataset <- wrap_expression(
  counts = fibroblast_reprogramming_treutlein$counts,
  expression = fibroblast_reprogramming_treutlein$expression
)
                               
ti_scfates_tree = create_ti_method_container("renjun0324/ti_scfates_tree:v0.1.0")
model = infer_trajectories(dataset_wrap, 
			    ti_scfates_tree(), 
			    parameters = list(tree_method="ppt"), # or epg
			    verbose = TRUE, return_verbose = TRUE, debug = FALSE)
model$model = map(model$model, add_cell_waypoints)
metric <- map_dfr(model$model,
                  dyneval::calculate_metrics,
                  dataset = dataset,
                  metrics = c("featureimp_wcor", "him",  "F1_branches", "correlation"))