TIGER main function

## Usage

```
tiger(
expr,
prior,
method = "VB",
TFexpressed = TRUE,
signed = TRUE,
baseline = TRUE,
psis_loo = FALSE,
seed = 123,
out_path = NULL,
out_size = 300,
a_sigma = 1,
b_sigma = 1,
a_alpha = 1,
b_alpha = 1,
sigmaZ = 10,
sigmaB = 1,
tol = 0.005
)
```

## Arguments

- expr
A normalized log-transformed gene expressison matrix. Rows are genes and columns are sampeles (cells).

- prior
A prior regulatory network in adjacency matrix format. Rows are TFs and columns target genes.

- method
Method used for Bayesian inference. "VB" or "MCMC". Defaults to "VB".

- TFexpressed
TF mRNA needs to be expressed or not. Defaults to TRUE.

- signed
Prior network is signed or not. Defaults to TRUE.

- baseline
Include baseline or not. Defaults to TRUE.

- psis_loo
Use pareto smoothed importance sampling leave-one-out cross validation to check model fitting or not. Defaults to FALSE.

- seed
Seed for reproducible results. Defaults to 123.

- out_path
(Optional) output path for CmdStanVB or CmdStanMCMC object. Defaults to NULL.

- out_size
Posterior sampling size. Default = 300.

- a_sigma
Hyperparameter of error term. Default = 1.

- b_sigma
Hyperparameter of error term. Default = 1.

- a_alpha
Hyperparameter of edge weight W. Default = 1.

- b_alpha
Hyperparameter of edge weight W. Default = 1.

- sigmaZ
Standard deviation of TF activity Z. Default = 10.

- sigmaB
Standard deviation of baseline term. Default = 1.

- tol
Convergence tolerance on ELBO.. Default = 0.005.

## Value

A TIGER list object. * W is the estimated regulatory network, but different from prior network, rows are genes and columns are TFs. * Z is the estimated TF activities, rows are TFs and columns are samples. * TF.name, TG.name, and sample.name are the used TFs, target genes and samples. * If psis_loo is TRUE, loocv is a table of psis_loo result for model checking. * If psis_loo is TRUE, elpd_loo is the Bayesian LOO estimate of the expected log pointwise predictive density, which can be used for Bayesian stacking to handle multi-modality later.