The 3D architecture of the pepper (Capsicum annum) genome and its relationship to function and evolution

The architecture of topologically associating domains (TADs) varies across plant genomes. Understanding the functional consequences of this diversity requires insights into the pattern, structure, and function of TADs. Here, we present a comprehensive investigation of the 3D genome organization of pepper (Capsicum annuum) and its association with gene expression and genomic variants. We report the first chromosome-scale long-read genome assembly of pepper and generate Hi-C contact maps for four tissues. The contact maps indicate that 3D structure varies somewhat across tissues, but generally the genome was segregated into subcompartments that were correlated with transcriptional state. In addition, chromosomes were almost continuously spanned by TADs, with the most prominent found in large genomic regions that were rich in retrotransposons. A substantial fraction of TAD boundaries were demarcated by chromatin loops, suggesting loop extrusion is a major mechanism for TAD formation; many of these loops were bordered by genes, especially in highly repetitive regions, resulting in gene clustering in three dimensional space. Integrated analysis of Hi-C profiles and transcriptomes showed that change in 3D chromatin structures (e.g. subcompartments, TADs, and loops) was not the primary mechanism contributing to differential gene expression between tissues, but chromatin structure does play a role in transcription stability. TAD boundaries were significantly enriched for breaks of synteny and depletion of sequence variation, suggesting that TADs constrain patterns of genome structural evolution in plants. Together, our work provides insights into principles of 3D genome folding in large plant genomes and its association with function and evolution.

Epiphany: predicting Hi-C contact maps from 1D epigenomic signals

Recent deep learning models that predict the Hi-C contact map from DNA sequence achieve promising accuracy but cannot generalize to new cell types and indeed do not capture cell-type-specific differences among training cell types. We propose Epiphany, a neural network to predict cell-type-specific Hi-C contact maps from five epigenomic tracks that are already available in hundreds of cell types and tissues: DNase I hypersensitive sites and ChIP-seq for CTCF, H3K27ac, H3K27me3, and H3K4me3. Epiphany uses 1D convolutional layers to learn local representations from the input tracks, a bidirectional long short-term memory (Bi-LSTM) layers to capture long term dependencies along the epigenome, as well as a generative adversarial network (GAN) architecture to encourage contact map realism. To improve the usability of predicted contact matrices, we trained and evaluated models using multiple normalization and matrix balancing techniques including KR, ICE, and HiC-DC+ Z-score and observed-over-expected count ratio. Epiphany is trained with a combination of MSE and adversarial (i.a., a GAN) loss to enhance its ability to produce realistic Hi-C contact maps for downstream analysis. Epiphany shows robust performance and generalization to held-out chromosomes within and across cell types and species, and its predicted contact matrices yield accurate TAD and significant interaction calls. At inference time, Epiphany can be used to study the contribution of specific epigenomic peaks to 3D architecture and to predict the structural changes caused by perturbations of epigenomic signals.

NucleoMap: a computational tool for identifying nucleosomes in ultra-high resolution contact maps

Although poorly positioned nucleosomes are ubiquitous in the prokaryote genome, they are difficult to identify with existing nucleosome identification methods. Recently available enhanced high-throughput chromatin conformation capture techniques such as Micro-C, DNase Hi-C, and Hi-CO characterize nucleosome-level chromatin proximity, probing the positions of mono-nucleosomes and the spacing between nucleosome pairs at the same time, enabling profiling of nucleosomes in poorly positioned regions. Here we develop a novel computational approach, NucleoMap, to identify nucleosome positioning from ultra-high resolution chromatin contact maps. By integrating nucleosome binding preferences, read density, and pairing information, NucleoMap precisely locates nucleosomes in both eukaryotic and prokaryotic genomes and outperforms existing nucleosome identification methods in sensitivity and specificity. We rigorously characterize genome-wide association in eukaryotes between the spatial organization of mono-nucleosomes and their corresponding histone modifications, protein binding activities, and higher-order chromatin functions. We also predict two tetra-nucleosome folding structures in human embryonic stem cells using machine learning methods and analysis their distribution at different structural and functional regions. Based on the identified nucleosomes, nucleosome contact maps are constructed, reflecting the inter-nucleosome distances and preserving the original data's contact distance profile.

A compendium of chromatin contact maps reflecting regulation by chromatin remodelers in budding yeast

We herein employ in situ Hi-C with an auxin-inducible degron (AID) system to examine the effect of chromatin remodeling on 3D genome organization in yeast. Eight selected ATP-dependent chromatin remodelers representing various subfamilies contribute to 3D genome organization differently. Among the studied remodelers, the temporary depletions of Chd1p, Swr1p, and Sth1p (a catalytic subunit of the Remodeling the Structure of Chromatin [RSC] complex) cause the most significant defects in intra-chromosomal contacts, and the regulatory roles of these three remodelers in 3D genome organization differ depending on the chromosomal context and cell cycle stage. Furthermore, even though Chd1p and Isw1p are known to share functional similarities/redundancies, their depletions lead to distinct effects on 3D structures. The RSC and cohesin complexes also differentially modulate 3D genome organization within chromosome arm regions, whereas RSC appears to support the function of cohesin in centromeric clustering at G2 phase. Our work suggests that the ATP-dependent chromatin remodelers control the 3D genome organization of yeast through their chromatin-remodeling activities.

Revisiting the Use of Structural Similarity Index in Hi-C

Identification of dynamic changes in chromatin conformation is a fundamental task in genetics. In 2020, Galan et al.1 presented CHESS (Comparison of Hi-C Experiments using Structural Similarity), a novel computational algorithm designed for systematic identification of structural differences in chromatin-contact maps. Using CHESS, the same group recently reported that chromatin organization is largely maintained across tissues during dorsoventral patterning of fruit fly embryos despite tissue-specific chromatin states and gene expression2. However, here we show that the primary outputs of CHESS–namely, the structural similarity index (SSIM) profiles–are nearly identical regardless of the input matrices, even when query and reference reads were shuffled to destroy any significant differences. This issue stems from the dominance of the regional counting noise arising from stochastic sampling in chromatin-contact maps, reflecting a fundamentally incorrect assumption of the CHESS algorithm. Therefore, biological interpretation of SSIM profiles generated by CHESS requires considerable caution.

Improving fragment-based ab initio protein structure assembly using low-accuracy contact-map predictions

Sequence-based contact prediction has shown considerable promise in assisting non-homologous structure modeling, but it often requires many homologous sequences and a sufficient number of correct contacts to achieve correct folds. Here, we developed a method, C-QUARK, that integrates multiple deep-learning and coevolution-based contact-maps to guide the replica-exchange Monte Carlo fragment assembly simulations. The method was tested on 247 non-redundant proteins, where C-QUARK could fold 75% of the cases with TM-scores (template-modeling scores) ≥0.5, which was 2.6 times more than that achieved by QUARK. For the 59 cases that had either low contact accuracy or few homologous sequences, C-QUARK correctly folded 6 times more proteins than other contact-based folding methods. C-QUARK was also tested on 64 free-modeling targets from the 13th CASP (critical assessment of protein structure prediction) experiment and had an average GDT_TS (global distance test) score that was 5% higher than the best CASP predictors. These data demonstrate, in a robust manner, the progress in modeling non-homologous protein structures using low-accuracy and sparse contact-map predictions.

Characterizing collaborative transcription regulation with a graph-based deep learning approach

Human epigenome and transcription activities have been characterized by a number of sequence-based deep learning approaches which only utilize the DNA sequences. However, transcription factors interact with each other, and their collaborative regulatory activities go beyond the linear DNA sequence. Therefore leveraging the informative 3D chromatin organization to investigate the collaborations among transcription factors is critical. We developed ECHO, a graph-based neural network, to predict chromatin features and characterize the collaboration among them by incorporating 3D chromatin organization from 200-bp high-resolution Micro-C contact maps. ECHO predicted 2,583 chromatin features with significantly higher average AUROC and AUPR than the best sequence-based model. We observed that chromatin contacts of different distances affected different types of chromatin features' prediction in diverse ways, suggesting complex and divergent collaborative regulatory mechanisms. Moreover, ECHO was interpretable via gradient-based attribution methods. The attributions on chromatin contacts identify important contacts relevant to chromatin features. The attributions on DNA sequences identify TF binding motifs and TF collaborative binding. Furthermore, combining the attributions on contacts and sequences reveals important sequence patterns in the neighborhood which are relevant to target sequence's chromatin feature prediction. The attribution results that reveal TF collaboration activities are provided on a website https://echo.dcmb.med.umich.edu/echo/.

Chromosomal-scale de novo genome assemblies of Cynomolgus Macaque and Common Marmoset

Cynomolgus macaque (Macaca fascicularis) and common marmoset (Callithrix jacchus) have been widely used in human biomedical research. Long-standing primate genome assemblies used the human genome as a reference for ordering and orienting the assembled fragments into chromosomes. Here we performed de novo genome assembly of these two species without any human genome-based bias observed in the genome assemblies released earlier. We assembled PacBio long reads, and the resultant contigs were scaffolded with Hi-C data, which were further refined based on Hi-C contact maps and alternate de novo assemblies. The assemblies achieved scaffold N50 lengths of 149 Mb and 137 Mb for cynomolgus macaque and common marmoset, respectively. The high fidelity of our assembly is also ascertained by BAC-end concordance in common marmoset. Our assembly of cynomolgus macaque outperformed all the available assemblies of this species in terms of contiguity. The chromosome-scale genome assemblies produced in this study are valuable resources for non-human primate models and provide an important baseline in human biomedical research.