Fig. 6: scCASE can effectively incorporate reference data to further promote enhancement performance.

a The graphical illustration of scCASER. scCASER builds upon scCASE by additionally incorporating reference data, which serves as the second input to the model. scCASER performs NMF on the reference data and incorporates the factorized projection matrix from the reference data. b Performance comparison between scCASER and scCASE. c The clustering performance of scCASER on the Spleen dataset with different numbers of self-reference pseudo-bulk samples. The dashed line represents the performance of scCASE method. Source data are provided as a Source Data file.