Analysis of the structural variability of topologically associated domains as revealed by Hi-C

Abstract

Three-dimensional chromosome structure plays an integral role in gene expression and regulation, replication timing, and other cellular processes. Topologically associated domains (TADs), building blocks of chromosome structure, are genomic regions with higher contact frequencies within the region than outside the region. A central question is the degree to which TADs are conserved or vary between conditions. We analyze 137 Hi-C samples from 9 studies under 3 measures to quantify the effects of various sources of biological and experimental variation. We observe significant variation in TAD sets between both non-replicate and replicate samples, and provide initial evidence that this variability does not come from genetic sequence differences. The effects of experimental protocol differences are also measured, demonstrating that samples can have protocol-specific structural changes, but that TADs are generally robust to lab-specific differences. This study represents a systematic quantification of key factors influencing comparisons of chromosome structure, suggesting significant variability and the potential for cell-type-specific structural features, which has previously not been systematically explored. The lack of observed influence of heredity and genetic differences on chromosome structure suggests that factors other than the genetic sequence are driving this structure, which plays an important role in human disease and cellular functioning.

SpectralTAD: an R package for defining a hierarchy of Topologically Associated Domains using spectral clustering

“…Expectedly, as the data resolution becomes lower, hence, becoming less sparse, consistency of all TAD callers increased, with SpectralTAD and TopDom performing the best. Previous work using data at 100kb resolution has shown that around 60-70% of boundaries are shared between primary and replicates (Dixon et al 2012;Rao et al 2014;Sauerwald et al 2018), and our results are in agreement with these observations. In summary, these results demonstrate robust performance of SpectralTAD in detecting consistent TAD boundaries across resolutions and replicates of Hi-C data.…”

TADCompare: An R Package for Differential and Temporal Analysis of Topologically Associated Domains

“…Previous studies have shown that the overlap between TAD boundaries in replicate data ranges from around 60 to 70% (Dixon et al, 2012;Rao et al, 2014;Sauerwald et al, 2020). Additionally, technical replicates have been shown to have a slightly higher proportion of shared TAD boundaries (∼65%) than biological replicates (∼60%) (Sauerwald et al, 2020). We have tested and confirmed these observations by showing that significantly more boundaries were non-differential in technical replicates than in biological replicates (73 vs. 65.7%).…”

“…Previous studies have shown that the overlap between TAD boundaries in replicate data ranges from around 60 to 70% (Dixon et al, 2012;Rao et al, 2014;Sauerwald et al, 2020). Additionally, technical replicates have been shown to have a slightly higher proportion of shared TAD boundaries (∼65%) than biological replicates (∼60%) (Sauerwald et al, 2020).…”

“…To our knowledge, there are only three methods that can be adapted for detecting changes in boundaries of interacting domains; the majority have been developed for the detection of TAD-specific boundary changes. Among the three methods, localtadsim (Sauerwald et al, 2020), HiCDB (Chen et al, 2018), and DiffTAD (Zaborowski and Wilczynski, 2016), none provide an intuitive, easy to use way of calling differential boundaries. Both localtadsim and DiffTAD are two-step procedures requiring separately defined TADs and comparing them using a commandline utility.…”

Capturing the complexity of topologically associating domains through multi-feature optimization

“…Because of the significant involvement of cohesin in TAD formation, two protein components of cohesin, RAD21 and SMC3, have also been shown to peak around TAD boundaries [24]. While TADs have been shown to vary between single cells [28], on the population level they are expected to remain fairly consistent between replicates [20]. On a fundamental level, TADs are defined by increased contact frequency within their interiors and somewhat depleted contact frequency between different TADs.…”