Bio-inspired crossover and a model of neuronal reprogramming

Abstract: Bio-inspired crossover:Natural genomes are affected by various kinds of mutations, manifesting as variations in the genome. This includes modifications to individual base pairs (single nucleotide polymorphism, SNP), but also short insertions and deletions (indels), and variations in the position (translocations) or multiplicity (copy number variations) of genetic sequences. While SNPs are historically the most studied, the other, structural variations are now known to affect a large fraction of the human genome, and can result in clinically relevant expressions in the phenotype. In biology, structural variations are also known to affect evolutionary dynamics of a population. For example, most genes originate through duplication, followed by divergent evolution of the two copies. However, in computer programs simulating evolution, structural variations are rarely included. Their absence makes it difficult to study the evolutionary consequences of translocation and copy number variation in simulated models. Including them may also have positive consequences for bio-inspired computation such as genetic algorithms. In this dissertation, a theoretical framework with algorithmic implementations is presented for sexual reproduction between linear genomes structures with structural variation, including permutations, and DNA-like strings.A model of neuronal reprogramming:In multicellular organisms, cells typically follow a hierarchical progression from embryonal stem cells to increasingly specialised cell types. However, by manipulating the expression of key genes, it is possible to reverse this specialisation, or convert directly between terminal cell types. This technology has been applied to convert human skin cells to neurons, with promising applications in disease modelling and regenerative medicine. Understanding the gene regulation network (GRN) governing this transition is crucial for improving the efficiency of this conversion in the future. In this dissertation, we integrated known interactions described in the literature into a holistic GRN model. We measured the activity of genes that are known to play pivotal roles during a reprogramming process, and compared these data to our modelling system. We found that the reprogramming process could be accurately modelled. The interaction between two genes, PTB and nPTB, played a different role from its usual interpretation in the literature, acting as a negative feedback loop.