scBatchCpp: Correct RNA-seq count matrix subject to batch effects by...

In tengfei-emory/scBatch: Batch Effect Correction Through Sample Distance Matrix Correction

Description

Use gradient descent algorithm to correct count matrix by linear transformation based on known sample matrix correction.

Usage

scBatchCpp(c, w, d, m, max, step, tol, derif, verbose)

Arguments

c	The original p*n batch effect data with n subjects and p RNA-seq measurements.
w	An initial n*n weight matrix to conduct linear transformation. An usual choice is the identity matrix if not specified.
d	The n*n distance matrix obtained by QuantNorm.
m	Number of groups to be divided for coordinate gradient descent. 1 < m <= n.
max	Maximum number of the iteration if the tolerance is not reached.
step	Step size of the gradient descent algorithm.
tol	Stopping criteria of the algorithm. The algorithm stops if the step size is smaller than tol.
derif	Function to compute gradient of the loss function. For Pearson distance, use scBatch::derif.
verbose	Whether output the loss function after each iteration.

Value

Returns the corrected count matrix.

Author(s)

Teng Fei. Email: [e-mail]

References

Fei and Yu (2020). scBatch: batch-effect correction of RNA-seq data through sample distance matrix adjustment. Bioinformatics, 36(10), 3115-3123. <https://doi.org/10.1093/bioinformatics/btaa097>

tengfei-emory/scBatch documentation built on June 6, 2021, 10:22 p.m.