Go to ICML 2025 Conference homepageReconstructing Cell Lineage Trees from Phenotypic Features with Metric Learning
Abstract: How a single fertilized cell gives rise to a complex array of specialized cell types in development is a central question in biology. The cells replicate to generate cell lineages and acquire differentiated characteristics through poorly understood molecular processes. A key approach to studying developmental processes is to infer the tree graph of cell lineage histories, which provides an analytical framework for dissecting individual cells' molecular decisions during replication and differentiation (i.e., acquisition of specialized traits). Although genetically engineered lineage-tracing methods have advanced the field, they are either infeasible or ethically constrained in many organisms. By contrast, modern single-cell technologies can measure high-content molecular profiles (*e.g.*, transcriptomes) in a wide range of biological systems. Here, we introduce *CellTreeQM*, a novel deep learning method based on transformer architectures that learns an embedding space with geometric properties optimized for tree-graph inference. By formulating the lineage reconstruction problem as tree-metric learning, we systematically explore weakly supervised training settings at different levels of information and present the *Cell Lineage Reconstruction Benchmark* to facilitate comprehensive evaluation. This benchmark includes (1) synthetic data modeled via Brownian motion with independent noise and spurious signals; (2) lineage-resolved single-cell RNA sequencing datasets. Experimental results show that *CellTreeQM* recovers lineage structures with minimal supervision and limited data, offering a scalable framework for uncovering cell lineage relationships. To our knowledge, this is the first method to cast cell lineage inference explicitly as a metric learning task, paving the way for future computational models aimed at uncovering the molecular dynamics of cell lineage. Code and benchmarks are available at: https://kuang-da.github.io/CellTreeQM-page
Lay Summary: How does a single fertilized cell become a whole organism? To understand this, scientists study the "family tree" of cells—how each cell gives rise to others during development. While genetic tools can track these relationships, they are often limited by technical or ethical constraints.
Our paper introduces a new AI-based method called CellTreeQM that reconstructs these cell family trees using molecular data from individual cells, such as gene activity (RNA levels), which can now be measured at scale thanks to modern biotechnology. Instead of relying on direct lineage labels, CellTreeQM learns how to position cells in a geometric space where distances reflect how closely related they are in the developmental process. The method uses a transformer-based deep learning model and learns from partial or indirect clues—like having only some parts of the cell tree or knowing which cells are grouped together. This level of supervision is often available in practice and relatively easy to extract from the data.
To rigorously test our method, we created a new benchmark with both simulated and real biological data. Our results show that CellTreeQM can uncover meaningful lineage relationships even with limited supervision, making it a practical tool for studying development in systems where traditional lineage tracing isn’t feasible. This work reframes lineage reconstruction as a metric learning problem and opens up new directions for using AI to study how cells develop and specialize.
Application-Driven Machine Learning: This submission is on Application-Driven Machine Learning.
Primary Area: Applications->Health / Medicine
Keywords: Metric learning, Cell lineage reconstruction, scRNA-seq, Representation learning
Submission Number: 5656
Loading