Teichlab
Something about relation of cell subsets between two datasets
Two datasets, D1 and D2 D1 contains cell type ct1 and ct2 D2 contains cell type ct3 and ct4
In my analysis, D1-ct1 corresponds to both D2-ct3 and D2-ct4, while D1-ct2 corresponds to UNRESOLVED. However, a meta-analysis suggests that D1-ct2 aligns with D2-ct4, which is consistent with biological expectations. The similarity between D1-ct1 and D2-ct4 is stronger than that between D1-ct2 and D2-ct4, complicating the interpretation.
The relational patterns—one-to-one, one-to-multiple, and multiple-to-one—do not fully capture this relationship. Specifically, D1-ct1 follows a one-to-multiple correspondence, while D2-ct4 exhibits a multiple-to-one pattern. Here, the classification of D1-ct2 as UNRESOLVED does not imply the discovery of a novel cell type, but rather reflects limitations in the matching process.
@niehu2018, thanks for using CellHint. Could you provide an example (with your real-world data and identified cell types rather than illustrative plots) to show the expected results from your data?
