Bonacker, Niklas (2022). Network inference and response prediction in biological systems. PhD thesis, Universität zu Köln.
|
PDF (PhD Thesis in Theoretical Physics of Niklas Bonacker)
Thesis.pdf - Published Version Download (3MB) | Preview |
Abstract
Anomalous biological information processing plays a crucial role in the formation and spread of cancer. Mathematical modelling and inference of signalling pathways is a central challenge in the field of cancer research [1]. We propose a stochastic model of gene regulation and solve our model within a Gaussian theory. We develop a maximum-likelihood estimate based on the Gaussian theory to infer regulatory interactions from steady state gene expression data. Within a simulated dataset, we compare our method to least squares fits, which are standard methods for gene regulatory network inference [7, 8]. Our estimate provides a more accurate network reconstruction in the regime of a sizeable stochastic contribution to the system dynamics. Based on perturbation experiments in the SK-MEL-133 melanoma cell line, we find that our maximum likelihood method leads to more accurate response predictions than least squares methods. High variability in patient response encumbers the clinical use of cancer immunotherapy. The understanding of determinants that drive immune response, resistance, and adverse side effects is a central scientific issue to move the field of cancer immunotherapy forward [2]. Several hypothetical response determinants are controversially discussed in the literature [3, 4, 5]. We search for patient-specific information about cancer immunotherapy response based on the tumour genome. Within the tumour genome, the contribution of frameshift mutations to immunotherapy response is less well studied [6]. Frameshift-derived peptides are very different from self-peptides and are a potential prime target for the immune system. Within our analysis, we focus, therefore, on frameshift-derived peptides. We find slight evidence that frameshift mutations are related to immunotherapy response. Nonetheless, our statistical analysis revealed that frameshift-derived peptides are not significantly associated with immunotherapy response. We find some evidence that a hidden factor, e.g. the mutation rate, increases both the number of unknown immunogenic mutations and the number of frameshift mutations. Still, there is no evidence that the frameshift mutations are causal for immunotherapy response.
Item Type: | Thesis (PhD thesis) | ||||||||||||||||
Translated title: |
|
||||||||||||||||
Translated abstract: |
|
||||||||||||||||
Creators: |
|
||||||||||||||||
Contributors: |
|
||||||||||||||||
URN: | urn:nbn:de:hbz:38-610877 | ||||||||||||||||
Date: | 12 April 2022 | ||||||||||||||||
Language: | English | ||||||||||||||||
Faculty: | Faculty of Mathematics and Natural Sciences | ||||||||||||||||
Divisions: | Faculty of Mathematics and Natural Sciences > Department of Physics > Institute for Theoretical Physics | ||||||||||||||||
Subjects: | Data processing Computer science Mathematics Physics Life sciences |
||||||||||||||||
Uncontrolled Keywords: |
|
||||||||||||||||
Date of oral exam: | 2 February 2022 | ||||||||||||||||
Referee: |
|
||||||||||||||||
Refereed: | Yes | ||||||||||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/61087 |
Downloads
Downloads per month over past year
Export
Actions (login required)
View Item |