Research: Previous Work

Subcellular gradients

Though morphogen gradients have been known experimentally for over 30 years (and theoretically predicted over 40 years ago by Lewis Wolpert), only recently have subcellular gradients been observed. Examples include the pom1p gradient in fission yeast (Padte et al. 2006) and gradients of phosoproteins (Brown and Kholodenko 1999, Chen et al., 2011). Subcellular gradients present new and interesting challenges when compared with "classical" morphogen gradients. Firstly, the particle number is much lower; perhaps fewer than 1000 signaling molecules in a cell. Second, the length scales are significantly shorter. Third, the relevant time scales can be very brief compared to morphogen readout in embryogenesis.

To investigate the effectiveness of subcellular gradients in determining positional precision I worked with Fred Chang at Columbia as part of his collaboration with Martin Howard. Fred's lab specialises in fission yeast and has made a number of important discoveries, such as recent work on how microtubules determine the cell division axis (Minc et al. 2011).

Along with Kally Pan, a PhD student in Fred's lab, we carefully quantified the pom1p gradient by measuring the pom1p profile in over 200 cells. We were able to quantify the fluctuations in the gradient from cell to cell. Furthermore, by using time lapse imaging, we have been able to quantify, for the first time, the intrinsic fluctuations in a spatial gradient.[9]

As a follow up to this work, we investigated the behaviour of Cdr2 during growth along with Nacho Flora (post-doc in Fred's lab). We found that Cdr2 plays an important role in cell cycle regulation: indeed, cell cycle progression was Cdr2 concentration dependent. Interestingly, during this work, we found that cells seem to "measure" their surface area as an important component in determining cell cycle progression.[12]

Tissue Mechanics

During my post-doc at EMBL-Heidelberg, I collaborated with Professor Maria Leptin and Matteo Rauzi on in toto imaging of gastrulation in Drosophila. I worked on software tools for handling the huge quantities of data produced for this project by multiview lightsheet microscopy. This project has recently been published in Nature Communications [16].

Morphogen gradients

in vivo measurements of morphogen dynamics

During my post-doc at EMBL-Heidelberg, I used lightsheet microscopy to quantify Bicoid gradient formation at very high temporal and spatial resolution. This was a challenging project due to the highly dynamic nature of the early embryo. Using a variety of techniques we have building up an unprecedented view of morphogen dynamics. This work is soon to be submitted for review.

Optimising morphogen spatial profiles to increase robustness

Initial work focused on understanding theoretically how the Bicoid gradient can be reliably interpreted by an embryo. By considering the effects of both embryo-to-embryo variation (such as differences in the amount of bcd mRNA inserted by mothers between eggs) and intrinsic fluctuations (caused by processes such as diffusion and protein degradation) on morphogen gradient precision it was found that the shape of gradients could be optimised to minimise the total effect of fluctuations. Interestingly, within relevant parameter regimes, exponentially decaying profiles were generally preferred, consistent with observation in vivo.[4]

in vivo testing of morphogen profile shape on readout

Along with Professor Howard, I worked with Professor Jun Ma in Cinncinnati to examine the Bicoid profile in nejire mutant embryos. This profile was altered compared to the wildtype - it is much better fitted by a power-law profile, rather than a simple exponential. The subsequent change in positional precision of the Hunchback boundary (the primary target of Bicoid) was correctly predicted from our theoretical considerations.[7]

Pre-steady-state intepretation of signaling molecules

Much of the original work on morphogen gradient interpretation assumed that the morphogen was in steady-state. However, experimental evidence increasingly suggests that morphogens may well be interpreted prior to obtaining steady-state (e.g. Sonic Hedgehog in vertebrate neural tube patterning). By considering both embryo-to-embryo and intrinsic fluctuations, simple criteria for when pre-steady-state interpretation of morphogen gradients was viable were developed.[5]

Lightsheet microscopy

Lightsheet microscopy enables whole embryo imaging while also keeping subcellular resolution. During my post-doc with Lars Hufnagel at EMBL-Heidelberg I worked on developing multiview lightsheet SPIM (MuVi-SPIM) [10]. In my lab we have built a similar lightsheet microscope that incorporates confocal lightsheet mode.

Synthetic enhancers

Synthetic enhancers provide a powerful tool for investigating transcription factor binding and cooperativity. Using the Drosophila mesoderm as a test system, the Furlong lab created a range of synthetic enhancers with different transcription factor binding motifs. Within this system, cooperativity between transcription factors played an important role in enhancer activity. I developed a simple fractional occupancy model replicated the data well and provided a powerful tool for understanding the results from multiple different synthetic enhancers.[11]


My Ph.D. was complete in the Rudolf Peierls Centre for Theoretical Physics at the University of Oxford. My research focus was on ordering in highly frustrated antiferromagnets. These systems are very interesting materials that have novel phase transitions at low temperatures. There has been recent focus on these materials due to the discovery of (effective) monopoles (Castelnovo et al., Nature 2008; S. T. Bramwell et al., Nature 2009).

bio physics
Pyrochlore frustrated antiferromagnet

It had been shown by John Chalker and Roderich Moessner (Moessner and Chalker, PRL 1998) that many frustrated antiferromagents should not order at finite temperature (due to the macroscopic number of available ground states). They also demonstrated that in particular cases an order-by-disorder transition can occur.

However, in experiments on frustrated antiferromagnets such as SCGO and ZnCr, clear ordering transitions were observed. In the former case, a transition to a spin freezing state was observed for sufficient impurity density. In the latter case, a Neel transition was possible under stress, but in the unstressed system ordering to a spin-glass-like state was observed at low levels of impurities. Importantly, the concentrations of impurities were significantly less than those typically observed in spin glasses.

Using Monte Carlo simulations (with parallelised code, running on up to 32 processors for a month), I investigated these transitions. I found that only very low levels of disorder in the bond lengths was sufficent to induce spin glass ordering [1],[6]. This work highlights the interesting physics that comes from considering spin interactions in constrained geometries.

Following on from this work, I looked to understand the Neel ordering transition and the crossover to spin-glass-like behaviour observed in ZnCr. This work combined phenomenological arguments with numerical results to derive a model that explained the observed experimental results. Though obviously a simplification of the real system (our approach was classical), the qualitiative agreement with experiment was good.[2]

I also worked on ordering out of Coulomb phases along with another PhD student, Tom Pickles.[3]

bio physics
Exploring optimisation of morpohgen gradient profile shape to ensure robust spatial boundary specification


bio physics
Pre-steady-state interpretation of morphogen gradients can result in robust specification of boundaries (as shown here for the gap gene network) but only under certain conditions


bio physics
To describe the penetrance (i.e. number of embryos displaying certain phenotype) of particular synthetic enhancers required the consideration of higher order interactions between transcription factors (i.e. direct interactions between the proteins, rather than simple DNA binding)


dev bio
Fission yeast cell division correlates strongly with surface area but not cell length or volume for cells with different geometr.


dev bio
Clustering model for the formation of a robust subcellular gradient in fission yeast


dev bio
Heat map of Bicoid intensity (x-axis, time (mins), y-axis, distance from anterior pole (µm), taken using a multi view lighsheet microscope