Bioinformatics Techniques to Analyze Dynamic Changes of 3D Genome

Project Summary
Three-dimensional (3D) folding of the genome plays fundamental roles in the regulation of transcription,
replication, DNA repair and many other biological processes. Facilitated by Hi-C and related techniques, it is
becoming clear that the eukaryotic genome folds at multiple genomic scales to form different types of 3D
architecture, including topologically associated domains (TADs) and stripes. Different physical patterns of
change may happen to a type of 3D architecture, e.g., a TAD may show change of overall connectivity, or split
into smaller TADs. Whereas the existence and functional importance of the genome’s 3D architecture is
increasingly recognized, analyzing its dynamic changes is currently a major challenge to biologists. The
community urgently needs novel bioinformatics techniques to define potential physical patterns of change for
each type of 3D architecture, to systematically detect all changes in the genome, and to statistically determine
the significance of each change. Our preliminary data strongly suggest that two physical patterns of change to
the genome’s 3D architecture -- TAD splittings and stripe strengthenings -- regulate cell identity transitions.
Accordingly, we propose to develop TADsplit and StripeDiff, two bioinformatics toolkits to systematically define
these and additionally physical patterns of change to TADs and stripes between samples. As a proof of
principle, we will utilize the new techniques to investigate 3D genome alterations during endothelial-to-
mesenchymal transition (EndMT), a cell identity transition that plays critical roles in both normal development
and many prevalent cardiovascular diseases. We will illustrate new mechanisms by which transcription factors
regulate genome’s 3D architectures to oppose EndMT. These investigations have the potential to better guide
the treatment of many diseases in which EndMT plays important roles. The novel bioinformatics techniques in
TADsplit and StripeDiff will enable researchers to investigate 3D genome changes in diverse biological models
of development and diseases.

Grant Number: 5R01GM138407-04
NIH Institute/Center: NIH
Principal Investigator: Kaifu Chen