Non-negative Matrix Factorization with phenotype and cell-cell similarity regularization.

Non-negative Matrix Factorization with phenotype and cell-cell similarity regularization, for identifing phenotype-associated cell states at different resolutions.

Usage

scAB.optimized(
  Object,
  K,
  alpha = 0.005,
  alpha_2 = 0.005,
  maxiter = 2000L,
  convergence_threshold = 1e-05
)

Arguments

Object

a scAB_data object

K

the rank of matrix factorization

alpha

Coefficient of phenotype regularization

alpha_2

Coefficient of cell-cell similarity regularization

maxiter

the maximum number of iterations

Value

a list with the submatrix and loss value

See also

Other scAB: DoscAB(), NMF.optimized(), create_scAB.v5(), findSubset.optimized(), select_K.optimized()