BGP: Branched Gaussian processes for identifying gene-specific branching dynamics in single cell data

Boukouvalas, Alexis; Hensman, James; Rattray, Magnus

doi:10.1101/166868

Cited by 5 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is not due to the varying developmental rates (such as in Experiment 3), then the method will try to explain the heterogeneity by the pseudotimes, causing an error in the method. Some pseudotime estimation methods are able do detect branches in the biological processes [4]. In such approach, one trajectory only takes into account data belonging to one branch, thus avoiding overfitting.…”

Section: Discussionmentioning

confidence: 99%

“…Firstly, we develop a method based on tracking the propagation of the (probability) distributions of cells over time, and finding a sparse matrix A that produces such propagation. The second paradigm is based on the so-called pseudotime concept [2]- [4]. The underlying idea in this concept is that cell dynamics are not identical through the population, and in particular, some cells develop faster than others.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Linear system identification from ensemble snapshot observations

Aalto

2019

2019 IEEE 58th Conference on Decision and Control (CDC)

View full text Add to dashboard Cite

Developments in transcriptomics techniques have caused a large demand in tailored computational methods for modelling gene expression dynamics from experimental data. Recently, so-called single-cell experiments have revolutionised genetic studies. These experiments yield gene expression data in single cell resolution for a large number of cells at a time. However, the cells are destroyed in the measurement process, and so the data consist of snapshots of an ensemble evolving over time, instead of time series. The problem studied in this article is how such data can be used in modelling gene regulatory dynamics. Two different paradigms are studied for linear system identification. The first is based on tracking the evolution of the distribution of cells over time. The second is based on the so-called pseudotime concept, identifying a common trajectory through the state space, along which cells propagate with different rates. Therefore, at any given time, the population contains cells in different stages of the trajectory. Resulting methods are compared in numerical experiments.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Linear system identification from ensemble snapshot observations

Aalto

2019

2019 IEEE 58th Conference on Decision and Control (CDC)

View full text Add to dashboard Cite

show abstract

“…Many methods exist to infer lineage relationships from single-cell expression profiles [10]. However, most retain no notion of uncertainty, do not infer differentially expressed genes in conjunction with reconstructing the tree, or require suitable normalization or dimensionality reduction beforehand [18][19][20][21][22][23][24][25][26][27][28]. Moreover, the latter preprocessing decisions are generally ad hoc rather than reflecting an explicit model of data collection-compounding the lack of a framework for uncertainty.…”

Section: Single-cell Trajectory Reconstructionmentioning

confidence: 99%

Reconstructing probabilistic trees of cellular differentiation from single-cell RNA-seq data

Shiffman¹,

Stephenson²,

Schiebinger³

et al. 2018

Preprint

View full text Add to dashboard Cite

Until recently, transcriptomics was limited to bulk RNA sequencing, obscuring the underlying expression patterns of individual cells in favor of a global average. Thanks to technological advances, we can now profile gene expression across thousands or millions of individual cells in parallel. This new type of data has led to the intriguing discovery that individual cell profiles can reflect the imprint of time or dynamic processes. However, synthesizing this information to reconstruct dynamic biological phenomena from data that are noisy, heterogenous, and sparse-and from processes that may unfold asynchronously-poses a complex computational and statistical challenge. Here, we develop a full generative model for probabilistically reconstructing trees of cellular differentiation from single-cell RNA-seq data. Specifically, we extend the framework of the classical Dirichlet diffusion tree to simultaneously infer branch topology and latent cell states along continuous trajectories over the full tree. In tandem, we construct a novel Markov chain Monte Carlo sampler that interleaves Metropolis-Hastings and message passing to leverage model structure for efficient inference. Finally, we demonstrate that these techniques can recover latent trajectories from simulated single-cell transcriptomes. While this work is motivated by cellular differentiation, we derive a tractable model that provides flexible densities for any data (coupled with an appropriate noise model) that arise from continuous evolution along a latent nonparametric tree.

show abstract

A comparison of single-cell trajectory inference methods: towards more accurate and robust tools

Saelens

Cannoodt

Todorov

et al. 2018

Preprint

View full text Add to dashboard Cite

Using single-cell -omics data, it is now possible to computationally order cells along trajectories, allowing the unbiased study of cellular dynamic processes. Since 2014, more than 50 trajectory inference methods have been developed, each with its own set of methodological characteristics. As a result, choosing a method to infer trajectories is often challenging, since a comprehensive assessment of the performance and robustness of each method is still lacking. In order to facilitate the comparison of the results of these methods to each other and to a gold standard, we developed a global framework to benchmark trajectory inference tools. Using this framework, we compared the trajectories from a total of 29 trajectory inference methods, on a large collection of real and synthetic datasets. We evaluate methods using several metrics, including accuracy of the inferred ordering, correctness of the network topology, code quality and user friendliness. We found that some methods, including Slingshot, TSCAN and Monocle DDRTree, clearly outperform other methods, although their performance depended on the type of trajectory present in the data. Based on our benchmarking results, we therefore developed a set of guidelines for method users. However, our analysis also indicated that there is still a lot of room for improvement, especially for methods detecting complex trajectory topologies. Our evaluation pipeline can therefore be used to spearhead the development of new scalable and more accurate methods, and is available at github.com/dynverse/dynverse. To our knowledge, this is the first comprehensive assessment of trajectory inference methods. For now, we exclusively evaluated the methods on their default parameters, but plan to add a detailed parameter tuning procedure in the future. We gladly welcome any discussion and feedback on key decisions made as part of this study, including the metrics used in the benchmark, the quality control checklist, and the implementation of the method wrappers. These discussions can be held at github.com/dynverse/dynverse/issues.

show abstract

BGP: Branched Gaussian processes for identifying gene-specific branching dynamics in single cell data

Cited by 5 publications

References 22 publications

Linear system identification from ensemble snapshot observations

Linear system identification from ensemble snapshot observations

Reconstructing probabilistic trees of cellular differentiation from single-cell RNA-seq data

A comparison of single-cell trajectory inference methods: towards more accurate and robust tools

Contact Info

Product

Resources

About