Sora Yoon, PhD, Atishay Jay, Golnaz Vahedi, PhD
Poster # 18
With the development of chromatin conformation capture techniques such as Hi-C and micro-C, diverse DNA folding structures have been elucidated, thereby presenting the potential to uncover the underlying causes of diseases through 3D genomics. In the field of medical biology, one of the primary research approaches is to identify (epi)genetic features that differentiate between two or more groups. However, the comparison between two groups using Hi-C or micro-C data has been challenging due to the noise inherent in the data. Therefore, in this study, we employed a noise-reduction strategy to reconstruct the network of chromatin interactions and subsequently compared two datasets by identifying sub-network modules in each dataset and comparing their module ranks. Following noise reduction, we discovered associations between network features such as connectivity, transitivity, and centrality nodes of each module and gene expression levels. Moreover, leveraging these network features, we confirmed that Hi-Cociety effectively identifies differentiating architectural features from the two datasets. Lastly, we established a connection between network features of single-cell level chromatin interaction data and Hi-C. In summary, Hi-Cociety serves as an R package for analyzing differences between two chromatin conformation capture datasets such as Hi-C and micro-C by employing noise filtering techniques and network-based approaches, enabling a comprehensive exploration of architectural disparities.
LIGHTNING TALK - 2023 MidAtlantic Bioinformatics Conference
Comments