Alignment of classical cell type definitions with molecular signatures

Determine the extent to which morphology- and phenotype-based cell type classifications correspond to molecular signatures measured in cells (e.g., transcriptomic, proteomic, epigenomic profiles), in order to clarify whether traditional cell type definitions align with distinct molecular states and signatures.

Background

Within the paper’s working definitions, the authors distinguish cell states, cell types, and cell fates. They note that cell types have traditionally been defined by morphology or phenotypic characteristics, while modern single-cell and multi-omic technologies increasingly use molecular signatures to characterize cellular identities.

The authors explicitly state that the concordance between classical, morphology-based cell type definitions and molecular signatures remains unresolved. This uncertainty has practical implications for Cell Atlas projects, mechanistic modeling, and the interpretation of single-cell data, and is linked to understanding sources and propagation of molecular noise.

References

The extent to which these classical definitions of cell types align with molecular signatures remains an open question, one that we will revisit throughout this review.

— Mapping, modeling, and reprogramming cell-fate decision making systems  (2412.00667 - Ham et al., 2024) in Subsubsection "Cell type" within Section "Working Definitions for Cell States, Types, and Fates"

Thus the proposed method suppresses dubious low-expression hits while retaining sensitivity to modest state changes (cycle or innate-response programs), though we cannot tell from this analysis alone whether those form stable clusters or transient states.

— Selective Inference for Deep Clustering in Latent Spaces  (2609.28756 - Mizui et al., 23 Sep 2026) in Appendix, Section 'PBMC: gene-level biological interpretation', subsection 'Gene-level tables and interpretation'