TASIC: determining branching models from time series single cell data

Rashid, Sabrina Mohd; Kotton, Darrell N.; Bar‐Joseph, Ziv

doi:10.1093/bioinformatics/btx173

Cited by 35 publications

(30 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Its successor, Monocle 2, uses a reversed graph embedding algorithm [5]. Other methods for scRNA-seq analysis use principal components analysis and hierarchical clustering (pcaReduce) [6], clustering based on approximate nearest neighbors (scmap) [7], biclustering (BackSpin) [8], a latent variable model (svLVM) [9], bifurcation analysis of a nearest neighbor graph (Wishbone) [10], a hidden Markov model coupled with probabilistic Kalman filtering (TASIC) [11], and a latent Dirichlet allocation model (cellTree) [12].…”

Section: Introductionmentioning

confidence: 99%

Unsupervised embedding of single-cell Hi-C data

Liu

Lin

Yardimci

2018

Preprint

View full text Add to dashboard Cite

Single-cell Hi-C (scHi-C) data promises to enable scientists to interrogate the 3D architecture of DNA in the nucleus of the cell, studying how this structure varies stochastically or along developmental or cell cycle axes. However, Hi-C data analysis requires methods that take into account the unique characteristics of this type of data. In this work, we explore whether methods that have been developed previously for the analysis of bulk Hi-C data can be applied to scHi-C data. In this work, we apply methods designed for analysis of bulk Hi-C data to scHi-C data in conjunction with unsupervised embedding. We find that one of these methods, HiCRep, when used in conjunction with multidimensional scaling (MDS), strongly outperforms three other methods, including a technique that has been used previously for scHi-C analysis. We also provide evidence that the HiCRep/MDS method is robust to extremely low per-cell sequencing depth, that this robustness is improved even further when high-coverage and low-coverage cells are projected together, and that the method can be used to jointly embed cells from multiple published datasets.

show abstract

Section: Introductionmentioning

confidence: 99%

Unsupervised embedding of single-cell Hi-C data

Liu

Lin

Yardimci

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…The use of branching probabilities leads to lower likelihood for cell assignments to later (more specific) paths in the branching tree. This is similar to prior probabilistic methods for reconstructing branching trajectories (Rashid et al, 2017). The idea here is that earlier stages are often less specific (higher entropy (Teschendorff and Enver, 2017), while later stages (representing specific fates) have a tighter expression profile.…”

Section: Cshmm Model Formulationmentioning

confidence: 88%

Continuous State HMMs for Modeling Time Series Single Cell RNA-Seq Data

Lin

Bar‐Joseph

2018

Preprint

Self Cite

View full text Add to dashboard Cite

Motivation: Methods for reconstructing developmental trajectories from time series single cell RNA-Seq (scRNA-Seq) data can be largely divided into two categories. The first, often referred to as pseudotime ordering methods, are deterministic and rely on dimensionality reduction followed by an ordering step. The second learns a probabilistic branching model to represent the developmental process. While both types have been successful, each suffers from shortcomings that can impact their accuracy. Results: We developed a new method based on continuous state HMMs (CSHMMs) for representing and modeling time series scRNA-Seq data. We define the CSHMM model and provide efficient learning and inference algorithms which allow the method to determine both the structure of the branching process and the assignment of cells to these branches. Analyzing several developmental single cell datasets we show that the CSHMM method accurately infers branching topology and correctly and continuously assign cells to paths, improving upon prior methods proposed for this task. Analysis of genes based on the continuous cell assignment identifies known and novel markers for different cell types.

show abstract

“…Since the SPRING analysis provides a low dimensional dynamic representation implying branching trajectories, we next sought to fully reconstruct these putative branching points to study their regulation and to characterize the set of TFs and signalling pathways associated with their potentially bifurcating fates. For this, we extended our previously developed computational method based on Hidden Markov models (Ding et al, 2018;Rashid et al, 2017) so that it can continuously assign cells along trajectories while still being able to infer regulators controlling branching events. The new method relies on Continuous State Hidden Markov Models (CSHMM).…”

Section: A Continuous Branching Network Model Learns Predicts and Mamentioning

confidence: 99%

Single-cell time-series mapping of cell fate trajectories reveals an expanded developmental potential for human PSC-derived distal lung progenitors

Hurley

Ding

Villacorta‐Martin

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

Alveolar epithelial type 2 cells (AEC2s) are the facultative progenitors responsible for maintaining lung alveoli throughout life, yet are difficult to access from patients for biomedical research or lung regeneration applications. Here we engineer AEC2s from human induced pluripotent stem cells (iPSCs) in vitro and use single cell RNA sequencing (scRNA-seq) to profile the detailed kinetics of their differentiation over time. We focus on both the desired target cells as well as those that appear to diverge to alternative endodermal fates. By combining scRNA-seq with lentiviral barcoding to trace differentiating clones, we reveal the bifurcating cell fate trajectories followed as primordial lung progenitors differentiate into mature AEC2s. We define the global transcriptomic signatures of primary developing human AEC2s from fetal through adult stages in order to identify the subset of in vitro differentiating cells that appear to recapitulate the path of in vivo development. In addition, we develop computational methods based on Continuous State Hidden Markov Models (CSHMM) to identify the precise timing and type of signals, such as over-exuberant Wnt responses, that induce some early multipotent NKX2-1+ progenitors to lose lung fate as they clonally diverge into a variety of non-lung endodermal lineages. Finally, we find that this initial developmental plasticity is regulatable via Wnt modulation, and subsides over time, ultimately resulting in iPSC-derived AEC2s that exhibit a stable phenotype and nearly limitless self-renewal capacity in vitro. Our methods and computational approaches can be widely applied to study and control directed differentiation, producing an inexhaustible supply of mature lineages, exemplified here by the generation of AEC2s.

show abstract

TASIC: determining branching models from time series single cell data

Abstract: Supplementary data are available at Bioinformatics online.

Cited by 35 publications

References 30 publications

Unsupervised embedding of single-cell Hi-C data

Unsupervised embedding of single-cell Hi-C data

Continuous State HMMs for Modeling Time Series Single Cell RNA-Seq Data

Single-cell time-series mapping of cell fate trajectories reveals an expanded developmental potential for human PSC-derived distal lung progenitors

Contact Info

Product

Resources

About