An interdisciplinary project with Miriam Gifford, Life Sciences, Warwick.

Context. Billions are spent each year on nitrogen fertiliser to sustain crop species in nitrogen-poor soils. Despite this massive investment much of the nitrogen is lost due to inefficient uptake by plants, polluting surrounding rivers with nitrates and ammonia. Fertiliser production is also a major contributor to greenhouse gases. There is thus a massive incentive to find a means to overcome fertilizer use. A possible solution is offered by the legumes which have evolved a symbiosis with a bacterium (Rhizobium) that can fix atmospheric nitrogen; these plants form nodules on their roots that ‘house’ the bacteria in exchange for nitrogen compounds (see figure). Transfer of this process to other plants is a grand challenge; we plan to make the first steps by developing a deep understanding of the process of nodulation, ascertaining the key genes that regulate it.

Nodules on medicago roots	GFP marked pericycle	GFP marked cortex

The project. A key hypothesis in the field is that nodulation in fact arose by co-opting an existing developmental blue-print, specifically that of lateral (side) root formation, an ancestral function of all plants. This project aims to examine this hypothesis and identify the key genes in nodulation by infering the regulatory network through comparing gene responses in root tissues to both nitrogen starvation and to the presence of the bacterium Rhizobium. By using time series and a comparative analysis we aim to gather sufficient data that the regulatory network can be identified. The project has two components:

Bioinformatic analysis (6 mths). A fundamental part of the project is the comparison between the legume Medicago and a model organism, Arabidopsis, which cannot form nodules. We will build bioinformatic data resources for Medicago, firstly, determining orthologs between Medicago and Arabidopsis using both sequence homology. Secondly, we will determine potential protein binding motifs in the Medicago genome, using existing sequence analysis software to perform de novo motif-finding.

Modelling and analysis (24 mths). Up to 16 time series will be generated from the experimental team, providing a global picture of the causal changes in gene expression in specific root tissues. We will perform initial analysis of this data, eg processing and clustering to determine the predominant patterns. Secondly, we will use the data to infer the nitrogen response regulatory network. There are a large variety of network inference methods. We will use a Bayesian approach to analyse this data, using both autoregressive models (linear and non linear) and factor models that integrate the motif data above. We will extend these models for both integration across experiments, including different genotypes (knocked out genes) and to incorporate a constrained topology variation model. We will further compare your results against possible functional classifications of the genes.

A fundamental experimental technique is the ability to mark particular cell types in the root, see figure. By isolating these tissues we can determine the gene expression in individual root tissues, and thus unravel differential expression and development signals in these tissues. Nodules for example, develop from the root cortex whilst lateral roots form in the pericycle.

The team. This is an interdisciplinary project involving the modellers Nigel Burroughs, Sascha Ott and the experimentalist Miriam Gifford (School of Life Sciences, Wellesborne campus). You will be expected to interact actively with the experimental team (1 PDRA, 2 PhD students). You will be supervised by Burroughs on modelling and networks, and Ott on bioinformatics.

Enquiries can be made to Nigel Burroughs, N dot J dot Burroughs at warwick dot ac dot uk.