scSLAT.model.batch.dual_pca

scSLAT.model.batch.dual_pca(X, Y, dim=50, singular=False, backend='sklearn', use_gpu=True)[source]

Dual PCA for batch correction

Parameters:

X (numpy.ndarray) – expr matrix 1 in shape of (cells, genes)
Y (numpy.ndarray) – expr matrix 2 in shape of (cells, genes)
dim (typing.Optional[int]) – dimension of embedding
singular (typing.Optional[bool]) – if multiply the singular value
backend (typing.Optional[str]) – backend to calculate singular value
use_gpu (typing.Optional[bool]) – if calculate in gpu

Return type:

typing.List[torch.Tensor]

Returns:

embd1, embd2 (Tensors of embedding)

References

Thanks Xin-Ming Tu for his [blog](https://xinmingtu.cn/blog/2022/CCA_dual_PCA/)