Identification and modelling of genetic features for genomic position profiling to support future approaches in gene therapy

Abstract

Genetic variability and mutations are a fundamental necessity for living organisms. In the context of evolution, both factors facilitate survival and enable the adaption of life to the environment. However, genetic variability also leads to risk factors for various medical conditions, and mutation can lead to divergent behavior, such as unsupervised division and proliferation of cells, known as cancer. With the cause being genetic, the most effective treatment and therapy is to modify the corresponding genetic sequence to repair or augmenting erroneous genes and silence risk factors, known colloquially as gene therapy. Besides applications in therapy, genetic modification is also a breakthrough technology for biotechnological engineering where cell cultures are utilized for drug production. Modifying genomes is a complex challenge because the correlation within the genome is neither fully understood, nor is the modification itself based on reliable mechanisms. Initial approaches to gene therapy in humans have shown that it is a potential game changer for many targets. However, immune reaction and insertional mutagenesis is a significant concern and universal application in therapy is only possible if major side-effects can be avoided. The foundation of gene therapy is the understanding of genomic function. Thus, in this dissertation, the aspect of transcript factor binding specificity is closely examined to gain new insights using a novel technology to analyze ChIP-Seq datasets. My approach is based on inferring binding models directly from the distributions of reads in relation to nearby sequences. This novel approach is capable of analyzing data sets that did not yield results using established methods. Furthermore, gene therapy relies on vectors that deliver genetic elements and insert them based on given targets. Therefore, a platform is presented to review insertional characteristics of genomic positions based on viral integration and transposases. Fundamental for the analysis is a mechanism to create computational background models that can be adapted for technological factors, as well as other known covariates. The applicability of the platform is shown in several publications that review genomic insertion preferences of delivery vectors.