Methods for mapping 3D chromosome architecture

Abstract

Several people have supported and helped me to during my PhD with guidance and advice, with contributions to the here presented work, but also with their friendship. First, I would like to thank my supervisor Ana for guiding me though my PhD and helping me manage three interesting but also challenging projects with ups and downs along the way. Thank you for lots of great advice and for helping me stay positive in difficult times.Many thanks to Uwe and Robert for our annual committee meetings, your time and advice, and my PhD committee for taking the time to review this thesis.I would like to thank Sasha (Alexander) for working with me on what sometimes seemed to be endless GAM optimisations, and for advice in all possible and impossible challenges that the wet lab provides. Many thanks go to all bioinformaticians that worked with me during my PhD; Ibai, Sasha (again), Ehsan, Christoph, Dominik, Tom, Rob, Mariano, and Markus. All of you are indispensable to making these projects a success, and you did not only do an amazing job but also managed to teach me a basic understanding of data analysis. Many thanks to Rob for all his prior work on GAM and for teaching me GAM when I joined the lab. Thank you, Enric, for joining our lab for the task of setting up an in-house WGA, your initiative and your work made this possible. Special thanks go to Gesa for helping me out when collecting nuclear profiles got too much for just one person, and for being a great friend. Without you the lab would have been only half the fun. I would also like to thank Izabela for her contribution to our final optimisations, which gave me a lot of positive energy when I really needed it, and everyone else in the Pombo lab for their help and/or friendship;

Probing multi-way chromatin interaction with hypergraph representation learning

“…Recent advances of ligation-free, genome-wide chromatin interaction mapping methods such as SPRITE (Quinodoz et al, 2018) and ChIA-Drop provide new perspectives on 3D genome and function by revealing multi-way contacts within the same nuclei (Kempfer and Pombo, 2019). However, computational methods that can fully utilize the potential of such multi-way chromatin interaction data remain underdeveloped.…”

“…However, one limitation of the proximity ligation based methods is that they capture interactions between genomic loci that are in close proximity to directly ligate and are unable to reveal chromatin interactions that are beyond the distance of direct ligation (Beagrie et al, 2017;Quinodoz et al, 2018). More importantly, most widely available Hi-C and ChIA-PET data only measure pairwise interactions and cannot delineate multiple chromatin loci that interact simultaneously in the same nucleus (Kempfer and Pombo, 2019).…”

FAN-C: A Feature-rich Framework for the Analysis and Visualisation of C data

“…Given the fundamental role that the correct organisation of chromatin in the nucleus plays for proper cell physiology, there is a growing need to integrate chromatin contact data in current stud-ies examining different aspects of gene and genome regulation. Different techniques have been developed to study chromatin conformation at the single cell or population level, in situ Hi-C being the primary method of choice for analysing chromatin conformation in cell populations (Rao et al 2014), reviewed in (Kempfer and Pombo 2019).…”

Comparison of the Hi-C, GAM and SPRITE methods by use of polymer models of chromatin

“…They provide a measure of the abundance of pairwise interactions, i.e., a Hi-C contact frequency map, by sequencing the ligation products of DNA fragments that are in close spatial proximity in the nucleus 13,15 . GAM probes 3D proximity of DNA sites by sequencing the genomic content of thin cryo-sectioned and laser micro-dissected slices from the nuclei of cells fixed in optimal preservation conditions 8,20 . Physically distant DNA sites are unlikely to co-segregate in the same thin slice, whereas physically proximal sites tend to co-segregate.…”