sim_seqdata: sim_seqdata

In DCLEAR: Distance Based Cell Lineage Reconstruction

sim_seqdata

Description

Generate singe cell barcode data set with tree shaped lineage information

Usage

sim_seqdata(
  sim_n = 200,
  m = 200,
  mu_d = 0.03,
  d = 15,
  n_s = 23,
  outcome_prob = NULL,
  p_d = 0.003
)

Arguments

sim_n	Number of cell samples to simulate.
m	Number of targets.
mu_d	Mutation rate. (a scalar or a vector)
d	Number of cell divisions.
n_s	Number of possible outcome states
outcome_prob	Outcome probability vector (default is NULL)
p_d	Dropout probability

Value

The result is a list containing two objects, 'seqs' and 'tree'. The 'seqs' is 'phyDat' object of 'sim_n' number of simulated barcodes corresponding to each cell, and The 'tree' is a 'phylo' object, a ground truth tree structure for the simulated data.

Author(s)

Il-Youp Kwak

Examples

library(DCLEAR)
library(phangorn)
library(ape)

set.seed(1)
mu_d1 = c( 30, 20, 10, 5, 5, 1, 0.01, 0.001)
mu_d1 = mu_d1/sum(mu_d1)
simn = 10 # number of cell samples
m = 10  ## number of targets
sD = sim_seqdata(sim_n = simn, m = m, mu_d = 0.03,
        d = 12, n_s = length(mu_d1), outcome_prob = mu_d1, p_d = 0.005 )
## RF score with hamming distance
D_hm = dist.hamming(sD$seqs)
tree_hm = NJ(D_hm)
RF.dist(tree_hm, sD$tree, normalize = TRUE)

## RF score with weighted hamming
InfoW = -log(mu_d1)
InfoW[1:2] = 1
InfoW[3:7] = 4.5
D_wh = dist_weighted_hamming(sD$seqs, InfoW, dropout=FALSE)
tree_wh = NJ(D_wh)
RF.dist(tree_wh, sD$tree, normalize = TRUE)

## RF score with weighted hamming, cosidering dropout situation
nfoW = -log(mu_d1)
InfoW[1] = 1
InfoW[2] = 12
InfoW[3:7] = 3
D_wh2 = dist_weighted_hamming(sD$seqs, InfoW, dropout = TRUE)
tree_wh2= NJ(D_wh2)
RF.dist(tree_wh2, sD$tree, normalize = TRUE)

DCLEAR documentation built on Sept. 14, 2023, 9:09 a.m.