Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

Significance When the human genome folds up inside the cell nucleus, it is spatially partitioned into numerous loops and contact domains. How these structures form is unknown. Here, we show that data from high-resolution spatial proximity maps are consistent with a model in which a complex, including the proteins CCCTC-binding factor (CTCF) and cohesin, mediates the formation of loops by a process of extrusion. Contact domains form as a byproduct of this process. The model accurately predicts how the genome will fold, using only information about the locations at which CTCF is bound. We demonstrate the ability to reengineer loops and domains in a predictable manner by creating highly targeted mutations, some as small as a single base pair, at CTCF sites. We recently used in situ Hi-C to create kilobase-resolution 3D maps of mammalian genomes. Here, we combine these maps with new Hi-C, microscopy, and genome-editing experiments to study the physical structure of chromatin fibers, domains, and loops. We find that the observed contact domains are inconsistent with the equilibrium state for an ordinary condensed polymer. Combining Hi-C data and novel mathematical theorems, we show that contact domains are also not consistent with a fractal globule. Instead, we use physical simulations to study two models of genome folding. In one, intermonomer attraction during polymer condensation leads to formation of an anisotropic “tension globule.” In the other, CCCTC-binding factor (CTCF) and cohesin act together to extrude unknotted loops during interphase. Both models are consistent with the observed contact domains and with the observation that contact domains tend to form inside loops. However, the extrusion model explains a far wider array of observations, such as why loops tend not to overlap and why the CTCF-binding motifs at pairs of loop anchors lie in the convergent orientation. Finally, we perform 13 genome-editing experiments examining the effect of altering CTCF-binding sites on chromatin folding. The convergent rule correctly predicts the affected loops in every case. Moreover, the extrusion model accurately predicts in silico the 3D maps resulting from each experiment using only the location of CTCF-binding sites in the WT. Thus, we show that it is possible to disrupt, restore, and move loops and domains using targeted mutations as small as a single base pair.

Chromatin extrusion explains key features of loop and domain formation in wild-type and engineered genomes

Resolving the 3D Landscape of Transcription-Linked Mammalian Chromatin Folding.

The establishment of cell types during development requires precise interactions between genes and distal regulatory sequences. We have a limited understanding of how these interactions look in three dimensions, vary across cell types in complex tissue, and relate to transcription. Here we describe optical reconstruction of chromatin architecture (ORCA), a method that can trace the DNA path in single cells with nanoscale accuracy and genomic resolution reaching two kilobases. We used ORCA to study a Hox gene cluster in cryosectioned Drosophila embryos and labelled around 30 RNA species in parallel. We identified cell-type-specific physical borders between active and Polycomb-repressed DNA, and unexpected Polycomb-independent borders. Deletion of Polycomb-independent borders led to ectopic enhancer-promoter contacts, aberrant gene expression, and developmental defects. Together, these results illustrate an approach for high-resolution, single-cell DNA domain analysis in vivo, identify domain structures that change with cell identity, and show that border elements contribute to the formation of physical domains in Drosophila.

/pdf/visualizing-dna-folding-and-rna-in-embryos-at-single-cell-rmna1c2bgg.pdf

Visualizing DNA folding and RNA in embryos at single-cell resolution

The human genome folds to create thousands of intervals, called "contact domains," that exhibit enhanced contact frequency within themselves. "Loop domains" form because of tethering between two loci-almost always bound by CTCF and cohesin-lying on the same chromosome. "Compartment domains" form when genomic intervals with similar histone marks co-segregate. Here, we explore the effects of degrading cohesin. All loop domains are eliminated, but neither compartment domains nor histone marks are affected. Loss of loop domains does not lead to widespread ectopic gene activation but does affect a significant minority of active genes. In particular, cohesin loss causes superenhancers to co-localize, forming hundreds of links within and across chromosomes and affecting the regulation of nearby genes. We then restore cohesin and monitor the re-formation of each loop. Although re-formation rates vary greatly, many megabase-sized loops recovered in under an hour, consistent with a model where loop extrusion is rapid.

Cohesin Loss Eliminates All Loop Domains

The regulatory specificity of enhancers and their interaction with gene promoters is thought to be controlled by their sequence and the binding of transcription factors. By studying Pitx1, a regulator of hindlimb development, we show that dynamic changes in chromatin conformation can restrict the activity of enhancers. Inconsistent with its hindlimb-restricted expression, Pitx1 is controlled by an enhancer (Pen) that shows activity in forelimbs and hindlimbs. By Capture Hi-C and three-dimensional modeling of the locus, we demonstrate that forelimbs and hindlimbs have fundamentally different chromatin configurations, whereby Pen and Pitx1 interact in hindlimbs and are physically separated in forelimbs. Structural variants can convert the inactive into the active conformation, thereby inducing Pitx1 misexpression in forelimbs, causing partial arm-to-leg transformation in mice and humans. Thus, tissue-specific three-dimensional chromatin conformation can contribute to enhancer activity and specificity in vivo and its disturbance can result in gene misexpression and disease.

Dynamic 3D chromatin architecture contributes to enhancer specificity and limb morphogenesis

Long-range gene regulation involves physical proximity between enhancers and promoters to generate precise patterns of gene expression in space and time. However, in some cases, proximity coincides with gene activation, whereas, in others, preformed topologies already exist before activation. In this study, we investigate the preformed configuration underlying the regulation of the Shh gene by its unique limb enhancer, the ZRS, in vivo during mouse development. Abrogating the constitutive transcription covering the ZRS region led to a shift within the Shh–ZRS contacts and a moderate reduction in Shh transcription. Deletion of the CTCF binding sites around the ZRS resulted in the loss of the Shh–ZRS preformed interaction and a 50% decrease in Shh expression but no phenotype, suggesting an additional, CTCF-independent mechanism of promoter–enhancer communication. This residual activity, however, was diminished by combining the loss of CTCF binding with a hypomorphic ZRS allele, resulting in severe Shh loss of function and digit agenesis. Our results indicate that the preformed chromatin structure of the Shh locus is sustained by multiple components and acts to reinforce enhancer–promoter communication for robust transcription.

https://www.pnas.org/content/pnas/116/25/12390.full.pdf

Preformed chromatin topology assists transcriptional robustness of Shh during limb development.

Genome-wide association studies (GWAS) have discovered thousands of risk loci for common, complex diseases, each of which could point to genes and gene programs that influence disease. For some diseases, it has been observed that GWAS signals converge on a smaller number of biological programs, and that this convergence can help to identify causal genes1–6. However, identifying such convergence remains challenging: each GWAS locus can have many candidate genes, each gene might act in one or more possible programs, and it remains unclear which programs might influence disease risk. Here, we developed a new approach to address this challenge, by creating unbiased maps to link disease variants to genes to programs (V2G2P) in a given cell type. We applied this approach to study the role of endothelial cells in the genetics of coronary artery disease (CAD). To link variants to genes, we constructed enhancer-gene maps using the Activity-by-Contact model7,8. To link genes to programs, we applied CRISPRi-Perturb-seq9–12 to knock down all expressed genes within ±500 Kb of 306 CAD GWAS signals13,14 and identify their effects on gene expression programs using single-cell RNA-sequencing. By combining these variant-to-gene and gene-to-program maps, we find that 43 of 306 CAD GWAS signals converge onto 5 gene programs linked to the cerebral cavernous malformations (CCM) pathway—which is known to coordinate transcriptional responses in endothelial cells15, but has not been previously linked to CAD risk. The strongest regulator of these programs is TLNRD1, which we show is a new CAD gene and novel regulator of the CCM pathway. TLNRD1 loss-of-function alters actin organization and barrier function in endothelial cells in vitro, and heart development in zebrafish in vivo. Together, our study identifies convergence of CAD risk loci into prioritized gene programs in endothelial cells, nominates new genes of potential therapeutic relevance for CAD, and demonstrates a generalizable strategy to connect disease variants to functions.

https://www.biorxiv.org/content/biorxiv/early/2022/11/04/2022.11.01.514606.full.pdf

Philine Guckelberger

Papers

Dynamic 3D chromatin architecture contributes to enhancer specificity and limb morphogenesis

Preformed chromatin topology assists transcriptional robustness of Shh during limb development.

Mapping the convergence of genes for coronary artery disease onto endothelial cell programs