run_magic: Do the magic!
In friedue/Rmagic: Markov Affinity-based Graph Imputation of Cells

1
2
3

run_magic(data, t_diffusion, lib_size_norm = TRUE, log_transform = FALSE,
  pseudo_count = 0.1, npca = 20, k = 30, ka = 10, epsilon = 1,
  rescale_percent = 0)

`data`	the data in some format
`t_diffusion`	diffusion time; suggested to start with 6 and increase it up to 12.
`lib_size_norm`	Default: TRUE.
`log_transform`	Default: FALSE.
`npca`	number of PCA components that should be used; default: 20.
`k`	Number of nearest neighbors to use when running MAGIC. Default: 30.
`ka`	kNN-autotune parameter for running MAGIC; default: 10
`epsilon`	a value for the standard deviation of the kernel. Default: 1. `epsilon = 0` is the uniform kernel.
`rescale_percent`	To which percentile should the data be re-scaled. Note: Do not set this higher than 0 if you also want to log-transform. Default: 0.
`pseudo-count`	A number indicating how much should be added to avoid log-transformation of zeros. Default: 0.1.