PhD Project Proposal:

 Oscillations matter: identifying networks of dynamically expressed genes

Supervisors:

Luisa Cutillo, School of Mathematics, UoL

David Westhead, School of Molecular and Cellular Biology, UoL

Potential External Collaborators:

1. Elli Marinopoulou and Nancy Papalopulu, University of Manchester (Papalopulu's Lab)
2.  MathWorks

• Measures the distribution of expression levels for each gene across a population of cells
• Allows to study new biological questions in which cell-specific changes in transcriptome are important, e.g. cell type identification, heterogeneity of cell responses, stochasticity of gene expression, inference of gene regulatory networks across the cells.

Focus on a single kind of omics:

Single Cell RNA sequencing

Focus on a specific kind of Networks:

genes co-oscillation networks

Why are oscillations important?

Oscillations in gene expression play a critical role in many biological developmental processes (eg. the circadian clock, cell cycle, neural development etc)

technologies for monitoring expression oscillations are limited also for well known processed like cell-cycle

Why the need for a model development?

Oscillators:

an alternative  view on gene expression

Example

two co-oscillating genes and the corresponding cell state

mRNA is collected at time T from (unsynchronised) cells in varying states

Background

OscoNet: inferring oscillatory gene networks

Luisa Cutillo, Alexis Boukouvalas, Elli Marinopoulou, Nancy Papalopulu & Magnus Rattray 

BMC Bioinformatics volume 21, Article number: 351 (2020) 

Identification of genes with oscillatory expression in glioblastoma: the paradigm of SOX2

Richard Zhiming Fu, Oliver Cottrell, Luisa Cutillo, Andrew Rowntree,Zsolt Zador,Heiko Wurdak, Nancy Papalopulu, and Elli Marinopoulou, Sci Rep. 2024; 14: 2123. Published online 2024 Jan 24.

• quiescent glioblastoma (GBM) stem cells (GSCs) play an important role in re-establishing the tumour

• protein oscillations can control the exit from quiescence of neural stem cells

• We identified a list of potential oscillators, including  SOX2

• We verified to oscillate in quiescent GSCs.

single cell RNA-seq analysis of adult and paediatric IDH-wildtype Glioblastomas (Neftel C, et al. Cell. 2019)

Starting dataset

Open problems

• Relax the assumption of homogenous cells population: assess the effect of population heterogeneity on the OscoNet performances
• Handle Data Sparsity within the model: assess how the inclusion of lowly-expressed genes affects the outcomes
• Integrate metadata (eg genes BP attributes) in the Network estimation and subsequent clustering
• Infer the period of oscillators using as prior the period of known oscillators (very few if not cell cycle!)
• Use the identified network of oscillatory genes to predict patients survival or tumour subtypes?

Questions?