CoChat

Reconstructing gene regulatory dynamics from high-dimensional single-cell snapshot data

Open access

The DOI matches its Crossref record; 37 of the 37 references it cites were found in a registry or an index.

Abstract

Motivation: High-dimensional single-cell snapshot data are becoming widespread in the systems biology community, as a mean to understand biological processes at the cellular level. However, as temporal information is lost with such data, mathematical models have been limited to capture only static features of the underlying cellular mechanisms. Results: Here, we present a modular framework which allows to recover the temporal behaviour from single-cell snapshot data and reverse engineer the dynamics of gene expression. The framework combines a dimensionality reduction method with a cell time-ordering algorithm to generate pseudo time-series observations. These are in turn used to learn transcriptional ODE models and do model selection on structural network features. We apply it on synthetic data and then on real hematopoietic stem cells data, to reconstruct gene expression dynamics during differentiation pathways and infer the structure of a key gene regulatory network. Availability and implementation: C++ and Matlab code available at https://www.helmholtz-muenchen.de/fileadmin/ICB/software/inferenceSnapshot.zip. Contact: fabian.theis@helmholtz-muenchen.de Supplementary information: Supplementary data are available at Bioinformatics online.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/bioinformatics/btv257

Cites

37 of 37 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1. Diagram of the framework. High-dimensional single-cell snapshot data (static data) are used as input in two paths (A). The first is a combination of a dimensionality reduction method [i.e. diffusion map algorithm (Coifman et al., 2005)] (B), clustering (C) and cell time-ordering [i.e. Wanderlust algorithm (Bendall et al., 2014)], which provides pseudo time-series (dynamic data) (D). Axis labels DC in the low-dimensional space, represent the first few diffusion components, i.e. eigenvectors. The second is a network inference algorithm [i.e. GENIE3 (Huynh-Thu et al., 2010)] generating a coarse GRN structure (E), which is used as prior knowledge during model selection. A likelihood function, which links transcriptional models M to the pseudo time-series data D, is used in a Bayesian framework to perform model selection and parameter estimation (F). Output is represented by a refined GRN structure (G) with corresponding posterior estimates of kinetic parameters (H)
Fig. 1. Diagram of the framework. High-dimensional single-cell snapshot data (static data) are used as input in two paths (A). The first is a combination of a dimensionality reduction method [i.e. diffusion map algorithm (Coifman et al., 2005)] (B), clustering (C) and cell time-ordering [i.e. Wanderlust algorithm (Bendall et al., 2014)], which provides pseudo time-series (dynamic data) (D). Axis labels DC in the low-dimensional space, represent the first few diffusion components, i.e. eigenvectors. The second is a network inference algorithm [i.e. GENIE3 (Huynh-Thu et al., 2010)] generating a coarse GRN structure (E), which is used as prior knowledge during model selection. A likelihood function, which links transcriptional models M to the pseudo time-series data D, is used in a Bayesian framework to perform model selection and parameter estimation (F). Output is represented by a refined GRN structure (G) with corresponding posterior estimates of kinetic parameters (H)CC BY-NC · PubMed Central
Fig. 2. (A) Comparison of reconstructed pseudo tim-series with time-series generated from true model. (B) Structure of incoherent type-I FFLs. (C) Three-dimensional embedding of snapshot data generated from OR-gate I1-FFL. Colours encode gZ expression levels. (D) Fit of AND-gate (solid green) and OR-gate (solid red) I1-FFL models to pseudo time-series data (black circles)
Fig. 2. (A) Comparison of reconstructed pseudo tim-series with time-series generated from true model. (B) Structure of incoherent type-I FFLs. (C) Three-dimensional embedding of snapshot data generated from OR-gate I1-FFL. Colours encode gZ expression levels. (D) Fit of AND-gate (solid green) and OR-gate (solid red) I1-FFL models to pseudo time-series data (black circles)CC BY-NC · PubMed Central
Fig. 3 Left: posterior parameter distributions, represented as univariate and bivariate marginal distributions, using OR-gate I1-FFL. True parameters used to simulate data are the following: α = 100, λ = 0.25, κ+ = 400, h+ = 20, κ− = 200, h− = 10. Right: ratio distribution obtained from posterior distributions over α and λ. True parameters ratio is given by α/λ = 400
Fig. 3 Left: posterior parameter distributions, represented as univariate and bivariate marginal distributions, using OR-gate I1-FFL. True parameters used to simulate data are the following: α = 100, λ = 0.25, κ+ = 400, h+ = 20, κ− = 200, h− = 10. Right: ratio distribution obtained from posterior distributions over α and λ. True parameters ratio is given by α/λ = 400CC BY-NC · PubMed Central
Fig. 4. Predictions generated by GRN inferred from snapshot data, under three different perturbation conditions. Noisy data generated from perturbed true OR-gate I1-FFL system (black dots), compared with predictions generated using inferred GRN and estimated parameters (solid red lines)
Fig. 4. Predictions generated by GRN inferred from snapshot data, under three different perturbation conditions. Noisy data generated from perturbed true OR-gate I1-FFL system (black dots), compared with predictions generated using inferred GRN and estimated parameters (solid red lines)CC BY-NC · PubMed Central
Fig. 5 Left: application of Wanderlust algorithm on one diffusion map’s branch from toggle switch simulated snapshot data. Colours encode time progression along the resulting trajectory. Right: reconstructed pseudo time-series, obtained by plotting gene expression values in time-ordered cells
Fig. 5 Left: application of Wanderlust algorithm on one diffusion map’s branch from toggle switch simulated snapshot data. Colours encode time progression along the resulting trajectory. Right: reconstructed pseudo time-series, obtained by plotting gene expression values in time-ordered cellsCC BY-NC · PubMed Central
Fig. 6 Left: model selection results for target gC. Model score in interval [0,1] represents a normalized AIC value: score =(AIC−m)/max(AIC−m), where m represents minimum AIC among all models. Right: fit of selected model (solid red) to gC pseudo time-series data reconstructed from two different branches (black circles)
Fig. 6 Left: model selection results for target gC. Model score in interval [0,1] represents a normalized AIC value: score =(AIC−m)/max(AIC−m), where m represents minimum AIC among all models. Right: fit of selected model (solid red) to gC pseudo time-series data reconstructed from two different branches (black circles)CC BY-NC · PubMed Central
Fig. 7. Hematopoietic data. (A) Differentiation lineages of HSCs. (B) Diffusion map of hematopoiesis single-cell snapshot data set. (C) GRN obtained after network inference on snapshot data
Fig. 7. Hematopoietic data. (A) Differentiation lineages of HSCs. (B) Diffusion map of hematopoiesis single-cell snapshot data set. (C) GRN obtained after network inference on snapshot dataCC BY-NC · PubMed Central
Fig. 8. Refined subnetwork using reconstructed pseudo time-series data (left) compared with experimental validated network in (Moignard et al., 2013) (right). Text label on each edge indicates the differentiation pathway used to predict that edge. In the experimental validated network, the known relationship between SCL and GATA2 (Moignard et al., 2013) is not shown; on the other hand, NFE2 is only predicted to be positively correlated with GATA2 in (Moignard et al., 2013)
Fig. 8. Refined subnetwork using reconstructed pseudo time-series data (left) compared with experimental validated network in (Moignard et al., 2013) (right). Text label on each edge indicates the differentiation pathway used to predict that edge. In the experimental validated network, the known relationship between SCL and GATA2 (Moignard et al., 2013) is not shown; on the other hand, NFE2 is only predicted to be positively correlated with GATA2 in (Moignard et al., 2013)CC BY-NC · PubMed Central
Fig. 9. Fitting of pseudo time-series from hematopoietic snapshot data with selected models. Observations from pseudo time-series (black circles), compared with model fitting (red lines)
Fig. 9. Fitting of pseudo time-series from hematopoietic snapshot data with selected models. Observations from pseudo time-series (black circles), compared with model fitting (red lines)CC BY-NC · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 1

  1. Jeffreys, 1961, Theory of Probability

Cited by

See all 155

Metadata

Authors
Ocone, A., Haghverdi, L., Mueller, N. S., & Theis, F. J.
Affiliations
Helmholtz Munich (DE)
Corresponding author
Theis, Fabian J.
Venue
Bioinformatics
Published
2015-06-10
Volume
31
Issue
12
Pages
i89-i96
ISSN
  • 1367-4803 (ISSN-L)
  • 1367-4811
OpenAlex
W2125258701
Cited by
155
References
36
Citations by year
2026: 7 · 2025: 6 · 2024: 12 · 2023: 16 · 2022: 11 · 2021: 8 · 2020: 20 · 2019: 22 · 2018: 17 · 2017: 19 · 2016: 14 · 2015: 3
FWCI
7.67
Type
journal article
Language
en
Open access
Locations
  • Bioinformatics (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
  • MDC Repository (Max-Delbrueck-Center for Molecular Medicine) (acceptedVersion)
  • HighWire Press Open Archive (submittedVersion)
  • Site cant be reached (submittedVersion, cc-by-nc)
  • PuSH - Publication Server of Helmholtz Zentrum München (submittedVersion)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
Keywords
gene regulatory network inference, gene expression dynamics, dimensionality reduction, model selection, hematopoietic stem cells, cell differentiation, temporal reconstruction
MeSH terms
Algorithms, Hematopoiesis, Hematopoietic Stem Cells, Kinetics, Models, Genetic, Gene Expression Profiling, Systems Biology, Gene Regulatory Networks, Single-Cell Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.