Learning Explicit Single-Cell Dynamics Using ODE Representations
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911483290451968 |
|---|---|
| author | von Bassewitz, Jan-Philipp Pervez, Adeel Fumero, Marco Robinson, Matthew Karaletsos, Theofanis Locatello, Francesco |
| author_facet | von Bassewitz, Jan-Philipp Pervez, Adeel Fumero, Marco Robinson, Matthew Karaletsos, Theofanis Locatello, Francesco |
| contents | Modeling the dynamics of cellular differentiation is fundamental to advancing the understanding and treatment of diseases associated with this process, such as cancer. With the rapid growth of single-cell datasets, this has also become a particularly promising and active domain for machine learning. Current state-of-the-art models, however, rely on computationally expensive optimal transport preprocessing and multi-stage training, while also not discovering explicit gene interactions. To address these challenges we propose Cell-Mechanistic Neural Networks (Cell-MNN), an encoder-decoder architecture whose latent representation is a locally linearized ODE governing the dynamics of cellular evolution from stem to tissue cells. Cell-MNN is fully end-to-end (besides a standard PCA pre-processing) and its ODE representation explicitly learns biologically consistent and interpretable gene interactions. Empirically, we show that Cell-MNN achieves competitive performance on single-cell benchmarks, surpasses state-of-the-art baselines in scaling to larger datasets and joint training across multiple datasets, while also learning interpretable gene interactions that we validate against the TRRUST database of gene interactions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_02903 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning Explicit Single-Cell Dynamics Using ODE Representations von Bassewitz, Jan-Philipp Pervez, Adeel Fumero, Marco Robinson, Matthew Karaletsos, Theofanis Locatello, Francesco Machine Learning Cell Behavior I.2.6; I.5.1; G.1.7; J.3 Modeling the dynamics of cellular differentiation is fundamental to advancing the understanding and treatment of diseases associated with this process, such as cancer. With the rapid growth of single-cell datasets, this has also become a particularly promising and active domain for machine learning. Current state-of-the-art models, however, rely on computationally expensive optimal transport preprocessing and multi-stage training, while also not discovering explicit gene interactions. To address these challenges we propose Cell-Mechanistic Neural Networks (Cell-MNN), an encoder-decoder architecture whose latent representation is a locally linearized ODE governing the dynamics of cellular evolution from stem to tissue cells. Cell-MNN is fully end-to-end (besides a standard PCA pre-processing) and its ODE representation explicitly learns biologically consistent and interpretable gene interactions. Empirically, we show that Cell-MNN achieves competitive performance on single-cell benchmarks, surpasses state-of-the-art baselines in scaling to larger datasets and joint training across multiple datasets, while also learning interpretable gene interactions that we validate against the TRRUST database of gene interactions. |
| title | Learning Explicit Single-Cell Dynamics Using ODE Representations |
| topic | Machine Learning Cell Behavior I.2.6; I.5.1; G.1.7; J.3 |
| url | https://arxiv.org/abs/2510.02903 |