Inferring single-cell transcriptomic dynamics with structured latent gene expression dynamics
Problem
Inferring developmental trajectories and gene expression dynamics from single-cell RNA sequencing data, where traditional RNA velocity assumptions fail with multiple lineages or time-dependent rates.
Why it matters
Understanding how cells differentiate is key to developmental biology, cancer and regenerative medicine research.
What the AI did
LatentVelo uses a variational autoencoder to embed cells in a latent space and neural ordinary differential equations to model differentiation dynamics and predict cell trajectories.
Results
not stated in source
Impact
The paper reports a method that infers latent trajectories for individual cells, but no quantitative benchmark is given in the abstract.
Record
| Venue | Cell Reports Methods |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Spencer Farrell, Madhav Mani, Sidhartha Goyal |
| Citations | 42 (OpenAlex, at import time) |
| References | 75 |
| Model or method | LatentVelo (VAE + neural ODEs) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1016/j.crmeth.2023.100581 |
| Open access | gold |
| Keywords | single-cell RNA sequencing, RNA velocity, variational autoencoder, neural ODE, trajectory inference, deep learning |
Importance rating
real but narrower advance
Trajectory inference from single-cell RNA-seq aids developmental biology, cancer and regenerative medicine research; broad field relevance but abstract gives no quantitative results or scale.