Inferring epigenetic dynamics from kin correlations

Sahand Hormoz, Nicolas Desprat, Boris I. ShraimanProceedings of the National Academy of Sciences 2015 · DOI

TL;DR

This paper introduces a novel method to infer the underlying dynamics of cellular phenotypes (like gene expression or protein production) by analyzing correlations between related cells in a population snapshot. This approach, inspired by physics' critical phenomena, allows researchers to understand how cells change over time and transmit traits, even without direct time-lapse observation of every cell.

What Problem Does This Paper Address?

Understanding how cells within a seemingly identical population (like stem cells or bacteria) develop diverse phenotypes and how these traits are passed down through generations is a fundamental challenge. While direct observation of cell lineages over time is ideal, it's often difficult or impossible, especially for large populations or when studying traits that require fixed-cell measurements. This paper tackles the problem of inferring these dynamic processes – how phenotypes change and are inherited – from static snapshots of cell populations, provided we know their family tree (genealogy).

Methods

The core of the paper is a statistical inference method based on a "minimal model" of epigenetic dynamics. This model assumes that a cell's phenotype is inherited from its parent with some probability, and that these transitions follow a Markovian process with detailed balance (meaning the process is time-reversible). The key innovation is relating the correlations between phenotypes of related cells (kin correlations) to the underlying transition probabilities.

They define "kin correlations" based on the genealogical distance between cells. By analyzing these correlations in a population snapshot, they can infer the parameters of the minimal model. They then extend this by introducing a more general model that allows for interactions between sister cells, which can be inferred from deviations in the kin correlations. This approach leverages concepts from conformal field theory, where higher-order correlations are constrained by lower-order ones, providing a powerful framework for inference.

Datasets

The primary dataset used for validation is experimental data on Pyoverdine (Pvd) dynamics in Pseudomonas aeruginosa colonies. This data consists of snapshots of Pvd fluorescence intensity in individual cells within microcolonies, along with their genealogical trees, collected over 9 generations.

Key Insights and Ideas

  • Phenotypic snapshots hold dynamic information: Even a static picture of a cell population, when combined with knowledge of their family relationships, can reveal how phenotypes change over time.
  • Conformal symmetry in biology: The correlations observed in phenotypic dynamics on genealogical trees exhibit a structure similar to critical phenomena in physics, specifically conformal symmetry. This suggests underlying universal principles governing these biological processes.
  • Inferring dynamics without direct observation: The developed method allows researchers to estimate transition rates and interaction parameters without needing to track every single cell's behavior over its entire lifespan.
  • Sister interactions matter: Deviations from the "minimal model" (which assumes independent inheritance) point to interactions between sister cells, influencing how phenotypes are passed down.

Key Results

  • The proposed inference method successfully extracts approximate probabilistic descriptions of phenotypic dynamics from kin correlations.
  • When applied to P. aeruginosa Pvd dynamics, the method's predictions for distant relatives closely matched experimental observations, validating the minimal model's core assumptions.
  • Analysis revealed that deviations from the minimal model could be explained by interactions between sister cells, which tend to reduce sharp phenotypic differences between them.
  • The study demonstrated that the structure of these correlations is remarkably constrained, with higher-order correlations being predictable from pair correlations, a hallmark of conformal symmetry.

Limitations and Future Work

  • Time scale limitations: The method is blind to phenotypic dynamics occurring on timescales shorter than a cell cycle or not transmitted epigenetically.
  • Spatial interactions: The current framework primarily focuses on genealogical relationships and doesn't fully capture direct spatial interactions between non-related cells, although it attempts to account for their indirect effects through sibling interactions.
  • Discretization: The phenotypic state space and time are discretized, which is an approximation.
  • Future Work: The authors suggest that this approach can be generalized to relax assumptions about detailed balance and synchronous division. They envision its application to understanding developmental programs, epigenetic states in stem cells, and cancer cell dynamics, particularly where lineage information is available but direct time-lapse imaging is challenging.

Summary updated

Abstract

Populations of isogenic embryonic stem cells or clonal bacteria often exhibit extensive phenotypic heterogeneity that arises from intrinsic stochastic dynamics of cells. The phenotypic state of a cell can be transmitted epigenetically in cell division, leading to correlations in the states of cells related by descent. The extent of these correlations is determined by the rates of transitions between the phenotypic states. Therefore, a snapshot of the phenotypes of a collection of cells with known genealogical structure contains information on phenotypic dynamics. Here, we use a model of phenotypic dynamics on a genealogical tree to define an inference method that allows extraction of an approximate probabilistic description of the dynamics from observed phenotype correlations as a function of the degree of kinship. The approach is tested and validated on the example of Pyoverdine dynamics in Pseudomonas aeruginosa colonies. Interestingly, we find that correlations among pairs and triples of distant relatives have a simple but nontrivial structure indicating that observed phenotypic dynamics on the genealogical tree is approximately conformal--a symmetry characteristic of critical behavior in physical systems. The proposed inference method is sufficiently general to be applied in any system where lineage information is available.
BibTeX
@inproceedings{hormoz2015inferring,
  author = {Sahand Hormoz and Nicolas Desprat and Boris I. Shraiman},
  title = {{Inferring epigenetic dynamics from kin correlations}},
  booktitle = {Proceedings of the National Academy of Sciences},
  year = {2015},
  doi = {10.1073/pnas.1504407112},
}
Figure