Run workflow to decompose tumor subclones
Usage
run_numbat(
count_mat,
lambdas_ref,
df_allele,
genome = "hg38",
out_dir = tempdir(),
max_iter = 2,
max_nni = 100,
t = 1e-05,
gamma = 20,
min_LLR = 5,
alpha = 1e-04,
eps = 1e-05,
max_entropy = 0.5,
init_k = 3,
min_cells = 50,
tau = 0.3,
nu = 1,
max_cost = ncol(count_mat) * tau,
n_cut = 0,
min_depth = 0,
common_diploid = TRUE,
min_overlap = 0.45,
ncores = 1,
ncores_nni = ncores,
random_init = FALSE,
segs_loh = NULL,
call_clonal_loh = FALSE,
verbose = TRUE,
diploid_chroms = NULL,
segs_consensus_fix = NULL,
use_loh = NULL,
min_genes = 10,
skip_nj = FALSE,
multi_allelic = TRUE,
p_multi = 1 - alpha,
plot = TRUE,
check_convergence = FALSE,
exclude_neu = TRUE
)Arguments
- count_mat
dgCMatrix Raw count matrices where rownames are genes and column names are cells
- lambdas_ref
matrix Either a named vector with gene names as names and normalized expression as values, or a matrix where rownames are genes and columns are pseudobulk names
- df_allele
dataframe Allele counts per cell, produced by preprocess_allele
- genome
character Genome version (hg38, hg19, or mm10)
- out_dir
string Output directory
- max_iter
integer Maximum number of iterations to run the phyologeny optimization
- max_nni
integer Maximum number of iterations to run NNI in the ML phylogeny inference
- t
numeric Transition probability
- gamma
numeric Dispersion parameter for the Beta-Binomial allele model
- min_LLR
numeric Minimum LLR to filter CNVs
- alpha
numeric P value cutoff for diploid finding
- eps
numeric Convergence threshold for ML tree search
- max_entropy
numeric Entropy threshold to filter CNVs
- init_k
integer Number of clusters in the initial clustering
- min_cells
integer Minimum number of cells to run HMM on
- tau
numeric Factor to determine max_cost as a function of the number of cells (0-1)
- nu
numeric Phase switch rate
- max_cost
numeric Likelihood threshold to collapse internal branches
- n_cut
integer Number of cuts on the phylogeny to define subclones
- min_depth
integer Minimum allele depth
- common_diploid
logical Whether to find common diploid regions in a group of peusdobulks
- min_overlap
numeric Minimum CNV overlap threshold
- ncores
integer Number of threads to use
- ncores_nni
integer Number of threads to use for NNI
- random_init
logical Whether to initiate phylogney using a random tree (internal use only)
- segs_loh
dataframe Segments of clonal LOH to be excluded
- call_clonal_loh
logical Whether to call segments with clonal LOH
- verbose
logical Verbosity
- diploid_chroms
vector Known diploid chromosomes
- segs_consensus_fix
dataframe Pre-determined segmentation of consensus CNVs
- use_loh
logical Whether to include LOH regions in the expression baseline
- min_genes
integer Minimum number of genes to call a segment
- skip_nj
logical Whether to skip NJ tree construction and only use UPGMA
- multi_allelic
logical Whether to call multi-allelic CNVs
- p_multi
numeric P value cutoff for calling multi-allelic CNVs
- plot
logical Whether to plot results
- check_convergence
logical Whether to terminate iterations based on consensus CNV convergence
- exclude_neu
logical Whether to exclude neutral segments from CNV retesting (internal use only)