Project descriptions

The IRTG 2466 "Network, exchange, and training program to understand plant resource allocation" invites applications for a PhD project

Information about application requirements and the online application form can be found under "Application & Finances"

Open Projects

Projects in Bioinformatics, Mathematics and related fields

Project ID: A3

Project title: Predictive Genotype-Phenotype Mapping using Deep Learning

Supervising PI: Markus Kollmann, Mathematical Modelling of Biological Systems

Research project: Advances in Machine Learning – in particular the use of Deep Neural Network – has led to unprecedented improvements on image classification tasks, superresolution, image generation, and predicting protein structure from multiple sequence alignments. In this project we will employ deep neural networks to predict how the life style of bacteria affects the use of synonymous codons. We expect to find differences in codon usage patterns especially between fast growing bacteria and bacteria that have been selected for efficient carbon metabolism. We are in particular interested how these patterns are related to pathway organization patterns. As the codon usage bias is strongly related to translational efficiency, the task is to predict the latter, which further involves accurate prediction of RNA secondary structures. Our approach makes use of autoregressive neural networks in combination with MCMC sampling methods to explore the likelihood function predicted from the data (~300M bacterial genomes). The predictions on translational efficiency and RNA folding will be validated by experiments carried out by the Axmann Group.

Required skills and qualifications: The successful candidate must have strong computational (python) and analytical skills, strong interest in life science problems, and be willing to dig deep into the deep learning literature. 

For specific questions regarding the PhD project, please contact Prof. Markus Kollmann: http://www.mathmodeling.hhu.de/en.html

Project ID: B2.2

Project title: A non-linear model for the growth of photosynthetic cells

Supervising PI: Martin Lercher, Computational Cell Biology

Research project: In many microbial organisms, natural selection has likely optimized the allocation of resources across the different cellular molecule species for fast growth. This likely also applies to cyanobacteria and unicellular algae. Accordingly, to simulate the growth and the basic physiology of photosynthetic organisms, one can formulate computer models that maximize the cellular growth rate in silico while accounting for important physical and biochemical constraints. This idea forms the basis of popular genome-scale modeling methods such as flux balance analysis (FBA), but also of more advanced methods such as resource balance analysis (RBA) and metabolism and expression (ME) models. However, all these methods ignore the concentrations of metabolites, which have been shown experimentally to strongly influence cellular performance.

Our group has developed a modeling scheme called growth balance analysis (GBA) that explicitly accounts for metabolite concentrations (Dourado & Lercher 2019, https://www.nature.com/articles/s41467-020-14751-w; see also https://www.biorxiv.org/content/10.1101/128009v1 and https://www.biorxiv.org/content/10.1101/802470v1). This scheme has not yet been applied to models of photosynthetic organisms - this is what we will do in the proposed project. We will start from available FBA models of cyanobacteria, which we will first reduce to a coarse-grained core model with lumped reactions. We will then integrate them with kinetic rate laws for the photosynthetic and metabolic reactions and a model of protein translation. Particular challenges will be the parameterisation of the reactions by comparison with experimental data obtained from literature sources and from collaborators at HHU, and the computational solution of the large non-linear optimisation problem. We aim to build a model that will not only allow us to understand the condition dependence of resource allocation in photosynthetic cells, but that will provide an accurate model to aid in the design of synthetic cells.

For specific questions regarding the PhD project, please contact Prof. Martin Lercher:https://www.cs.hhu.de/en/research-groups/computational-cell-biology/our-team/team.html

 

 

iGRAD-Plant Coordinator

Dr. Petra Fackendahl

Institute of Plant Biochemistry
Universitätsstr. 1
Gebäude: 26.13.
Etage/Raum: U1.78
40225 Düsseldorf
Tel.: +49 211 81-10588
Verantwortlich für den Inhalt: E-Mail sendenDr. Petra Fackendahl