Today we share some exciting news: a new paper is out today in the Science journal by DNA Zoo and collaborators!
In the manuscript, titled "3D genomics across the tree of life reveals condensin II as a determinant of architecture type" we use the DNA Zoo data to explore the nuclear architecture for 27 species across the eukaryotic tree of life.
We show that all of the chromosomal architectures observed across the species fall into one of just two types: those chromosome territory-like (with chromosomes occupying distinct nuclear subvolumes) and those Rabl-like (resembling the arrangement observed during cell division, with centromeres and/or telomeres clustered and chromosomes often "folded" along the telomere-to-centromere axis). We show that Rabl-like architecture in species is often associated with disruption of condensin II, a protein known to be responsible for lengthwise compaction of chromosomes.
We perform an experiment (led by our colleagues Claire Hoencamp and Benjamin Rowland at the Netherlands Cancer Institute) to disrupt condensin II in human cells. This transforms a territorial chromosome architecture typical of human cells into Rabl. We follow up with some physical simulations, led by our collaborators Sumitabha Brahmachari and José Onuchic at the Center for Theoretical Biological Physics at Rice University, suggesting that condensin II disruption results in long and floppy chromosomes that cannot generate mechanical tensions enough to disrupt the Rabl arrangement "inherited" from cell division. Read more about this in the manuscript and in the joint Baylor College of Medicine and Rice University press release.
As part of the effort described in the manuscript, we publish 17 new chromosome-length genome assemblies. We include the links to the assembly pages for those previously published and those shared today below.
We also generate chromosome-length haploblocks for 7 non-human species using a new Hi-C-based phaser, now part of 3D-DNA. The phaser uses a list of deduplicated Hi-C contacts, as generated by the Juicer pipeline (Durand, Shamim et al., 2016), to phase variants encoded in a vcf (variant call format) file. It plays well with prephased data, e.g. from linked reads, long reads or population data, and generates phasing contact maps to validate the results that can be further polished in Juicebox Assembly Tools. So, JBAT can now not only help with genome assembly, but also assists with phasing!
We explore the coverage requirements for whole chromosome phasing, and show that Hi-C based phasing works in most species, including human. However, we show that in species where the homologs are not separate during interphase, like Drosophila melanogaster, this method cannot be used. This highlights how the principles of genome assembly can vary across different complex eukaryotes.
We will write about the phasing piece separately, and in the meantime, please stay tuned for a few blog posts highlighting some of the newly shared genome assemblies in the next week or so!