High-throughput Chromosome Conformation Capture (3C) experiments, provide detailed three-dimensional (3D) information about genome organization. Specially, Hi-C, a 3C derivative, has become the standard technique to investigate the 3D chromatin structure, and its functional implication into cell fate determination. However, the correct bioinformatic analysis and interpretation of this data is still an active field of development.
In this thesis, we explore the ability of CTCF to form chromatin loops and their epigenetic signature, by developing Meta-Waffle, an artificial neural network to classify structural patterns without any prior information. This classification, was used to generate a convolutional neural network to de novo detect chromatin loops from Hi-C contact matrices, called LOOPbit.
We also present CHESS, a bioinformatic tool for the comparison of chromatin contact maps and differential 3D feature extraction, such as Topologically Associating Domains, stripes or loops.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados