UniTVelo: temporally unified RNA velocity reinforces single-cell trajectory inference

Gao, Mingze; Qiao, Chen; Huang, Yuanhua

doi:10.1101/2022.04.27.489808

Cited by 8 publications

(21 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We primarily compare our model with scVelo dynamical and stochastic modes [11]. Comparisons with the default settings of UniTVelo unified and independent modes [17], and DeepVelo [18] are shown in the supplemental. Note: scVelo stochastic mode is an updated version of the Velocyto steady-state model, where regression is done for the first 2 moments of the dynamics instead of just the mean [11].…”

Section: Resultsmentioning

confidence: 99%

“…When ground truth velocities are not known, we use known cell-type transitions with the Cross-Boundary Direct-edness metric [16, 17]. where s i, j = ( s i – s j )/sign( s i – s j ) and is the neighborhood of cell i .…”

Section: Methodsmentioning

confidence: 99%

“…For real data, we use datasets where the direction of differentiation is known. To quantitatively evaluate models, we use a modified version of the Cross-Boundary Directedness score used for UniTVelo and VeloAE [16,17]. This score measures the velocity direction of cells on the boundary between two cell-types.…”

Section: A Evaluationmentioning

confidence: 99%

“…VeloAE [16] develops an extension of the steady-state model by projecting high-dimensional noisy unspliced and spliced vectors onto a lower dimension latent space with an autoencoder, and then uses the steady-state linear model to estimate velocities on this latent space. UniTVelo [17] allows for a more general form of spliced RNA dynamics which relaxes assumptions on the transcription rate, includes a unified celldependent latent time (rather than per gene as in scVelo), and introduces a method of dealing with low signal-to noise in unspliced reads. DeepVelo [18] uses a deep neural network to estimate cell-specific kinetic rate parameters to model variable and lineage-dependent kinetics.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Inferring single-cell transcriptomic dynamics with structured latent gene expression dynamics

Farrell

Mani

Goyal

2022

Preprint

View full text Add to dashboard Cite

RNA velocity provides directional information for trajectory inference from single-cell RNA-sequencing data. Traditional approaches to computing RNA velocity rely on strict assumptions about the equations describing transcription of unspliced RNA and splicing of unspliced RNA into spliced RNA. This results in issues in scenarios where these assumptions are violated, such as multiple lineages with distinct dynamics and time-dependent rates. In this work we present ""LatentVelo"", a novel approach to computing a low-dimensional representation of RNA velocity with deep learning. Our approach embeds cells into a latent space with a variational auto-encoder, and describes differentiation dynamics on this latent space with neural ordinary differential equations. These more general dynamics enable accurate trajectory inference, and the latent space approach enables the generation of dynamics-based latent embeddings of cell states and batch correction of cell states and of RNA velocity. By performing simultaneous batch correction of RNA velocity dynamics and gene expression, we also outperform traditional batch correction methods that only consider gene expression. The flexible structure of the model enables modelling a variety of regulatory structures and multi-omic data, or incorporating additional information such as cell-type annotations or experimental metadata to improve the latent embedding. LatentVelo will be available at https://github.com/Spencerfar/LatentVelo.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: A Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Inferring single-cell transcriptomic dynamics with structured latent gene expression dynamics

Farrell

Mani

Goyal

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…A vector field is created based on the displacements between the future state and the current state in the spliced matrix space. Gao et al (2022) proposes a revised high-dimensional RNA velocities estimation model UniTVelo which imposes a gene-shared cell latent time to circumvent the independent estimation issue of other RNA velocity estimations approaches.…”

Section: Introductionmentioning

confidence: 99%

Pumping the brakes on RNA velocity – understanding and interpreting RNA velocity estimates

Stein-O’Brien

Boukas

Goff

et al. 2022

Preprint

View full text Add to dashboard Cite

RNA velocity analysis of single cells promises to predict temporal dynamics from gene expression. Indeed, in many systems, it has been observed that RNA velocity produces a vector field which qualitatively reflects known features of the system. Despite this observation, the limitations of RNA velocity estimates are poorly understood. Using real data and simulations, we dissect the impact of different steps in the RNA velocity workflow on the estimated vector field. We find that the process of mapping RNA velocity estimates into a low-dimensional representation, such as those produced by UMAP, has a large impact on the result. The RNA velocity vector field is strongly dependent on the k-NN graph of the data. This dependence lead to large estimator errors when the k-NN graph is not a faithful representation of the true data structure; a feature which cannot be known for most real datasets. Finally, we establish that RNA velocity does not estimate expression speed neither at the gene nor at the cellular level. We propose that RNA velocity is best considered a smoothed interpolation of the observed k-NN structure, as opposed to an extrapolation of future cellular states, and that the use of RNA velocity as a validation of latent space embedding structures is circular.

show abstract

Organoid-based single-cell spatiotemporal gene expression landscape of human embryonic development and hematopoiesis

Chao

Xiang

Xiao

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Single-cell level characterization of embryonic development is a major benchmark of human developmental biology. Spatiotemporal analysis of stem-cell-derived embryos offers conceptual and technical advances in the field. Here, we defined the single-cell spatiotemporal gene expression landscape of human embryonic development with stem-cell-derived organoids. We established the human embryonic organoid (HEMO) from expanded potential stem cells and achieved both embryonic and extraembryonic tissues in the same organoid. Time-series single-cell RNA sequencing paired with single-cell resolution spatial revealed human embryonic development signatures such as extraembryonic placenta, yolk sac hematopoiesis neural crest, blood vessels, and cardiac mesoderm. Hematopoietic tissues eventually predominated HEMO with erythropoiesis, mekagaryopiesis, and myelopoiesis. Cell-cell communication network analysis demonstrated that trophoblast-like tissues supplied WNT signaling in neural crest cells to facilitate maturation and migration. Single-cell resolution spatial transcriptomics defined the yolk sac erythro-megakaryopoietic niche. Vitronectin-integrin signaling, a major contributor to megakaryocyte maturation, was predominant in the yolk sac niche in HEMO and to human fetal samples. Overall, our study advances the spatiotemporal analysis of human embryonic development in stem-cell-derived organoids.

show abstract

UniTVelo: temporally unified RNA velocity reinforces single-cell trajectory inference

Cited by 8 publications

References 42 publications

Inferring single-cell transcriptomic dynamics with structured latent gene expression dynamics

Inferring single-cell transcriptomic dynamics with structured latent gene expression dynamics

Pumping the brakes on RNA velocity – understanding and interpreting RNA velocity estimates

Organoid-based single-cell spatiotemporal gene expression landscape of human embryonic development and hematopoiesis

Contact Info

Product

Resources

About