assemble_partition

Assembles a partition for a given cell dataset (CDS) by fitting wild-type (WT) and mutant (MT) models, constructing state transition graphs, and assessing perturbation effects.

assemble_partition(
  cds,
  sample_group,
  cell_group,
  partition_name = NULL,
  main_model_formula_str = NULL,
  start_time = 18,
  stop_time = 72,
  interval_col = "timepoint",
  nuisance_model_formula_str = "~1",
  ctrl_ids = NULL,
  mt_ids = NULL,
  sparsity_factor = 0.01,
  perturbation_col = "gene_target",
  batch_col = "expt",
  max_num_cells = NULL,
  verbose = FALSE,
  keep_ccs = TRUE,
  num_threads = 1,
  backend = "nlopt",
  q_val = 0.1,
  vhat_method = "bootstrap",
  num_bootstraps = 10,
  newdata = tibble::tibble(),
  edge_allowlist = NULL,
  min_lfc = 0,
  links_between_components = c("ctp", "none", "strongest-pcor", "strong-pcor"),
  log_abund_detection_thresh = -5,
  batches_excluded_from_assembly = c(),
  component_col = "partition",
  embryo_size_factors = NULL
)

Arguments

• cds
  CDS object
  Cell dataset to analyze.

• sample_group
  string
  Column name in CDS for sample groups.

• cell_group
  string
  Column name in CDS for cell groups.

• partition_name
  string
  Name of the partition. Default is NULL.

• main_model_formula_str
  string
  Main model formula. Default is NULL.

• start_time
  numeric
  Start time for analysis. Default is 18.

• stop_time
  numeric
  Stop time for analysis. Default is 72.

• interval_col
  string
  Column name for time intervals. Default is "timepoint".

• nuisance_model_formula_str
  string
  Nuisance model formula. Default is "~1".

• ctrl_ids
  vector
  Control IDs. Default is NULL.

• mt_ids
  vector
  Mutant IDs. Default is NULL.

• sparsity_factor
  numeric
  Sparsity factor. Default is 0.01.

• perturbation_col
  string
  Perturbation column name. Default is "gene_target".

• batch_col
  string
  Batch column name. Default is "expt".

• max_num_cells
  numeric
  Max number of cells to include. Default is NULL.

• verbose
  logical
  Print verbose output. Default is FALSE.

• keep_ccs
  logical
  Keep connected components. Default is TRUE.

• num_threads
  numeric
  Number of threads to use. Default is 1.

• backend
  string
  Optimization backend. Default is "nlopt".

• q_val
  numeric
  Q-value threshold for perturbation significance. Default is 0.1.

• vhat_method
  string
  Method for variance estimation. Default is "bootstrap".

• num_bootstraps
  numeric
  Number of bootstraps. Default is 10.

• newdata
  tibble
  New data for predictions. Default is an empty tibble.

• edge_allowlist
  list
  Allowed edges in graph. Default is NULL.

• min_lfc
  numeric
  Minimum log fold change. Default is 0.

• links_between_components
  character vector
  Types of links between components. Default is c("ctp", "none", "strongest-pcor", "strong-pcor").

• log_abund_detection_thresh
  numeric
  Log abundance detection threshold. Default is -5.

• batches_excluded_from_assembly
  vector
  Batches to exclude. Default is empty.

• component_col
  string
  Column name for components. Default is "partition".

• embryo_size_factors
  vector
  Size factors for embryos. Default is NULL.

Value

A tibble containing the results of the partition assembly, including WT and MT graphs, perturbation effects, and state graph plots.

Details

The function performs the following steps:

• Prepares the CDS by adding subassembly group and cell state information.
• Fits a wild-type model and assembles a WT state transition graph.
• Fits mutant models and assembles MT state transition graphs.
• Assesses perturbation effects and constructs annotated graphs.
• Handles errors gracefully and returns NA for failed steps.

Examples

results <- assemble_partition(
  cds = my_cds,
  sample_group = "sample",
  cell_group = "cell_type",
  partition_name = "partition_1",
  main_model_formula_str = "~timepoint",
  start_time = 18,
  stop_time = 72
)