Nuclear morphology is shaped by loop-extrusion programs.

It is well established that neutrophils adopt malleable polymorphonuclear shapes to migrate through narrow interstitial tissue spaces1-3. However, how polymorphonuclear structures are assembled remains unknown4. Here we show that in neutrophil progenitors, halting loop extrusion-a motor-powered process that generates DNA loops by pulling in chromatin5-leads to the assembly of polymorphonuclear genomes. Specifically, we found that in mononuclear neutrophil progenitors, acute depletion of the loop-extrusion loading factor nipped-B-like protein (NIPBL) induced the assembly of horseshoe, banded, ringed and hypersegmented nuclear structures and led to a reduction in nuclear volume, mirroring what is observed during the differentiation of neutrophils. Depletion of NIPBL also induced cell-cycle arrest, activated a neutrophil-specific gene program and conditioned a loss of interactions across topologically associating domains to generate a chromatin architecture that resembled that of differentiated neutrophils. Removing NIPBL resulted in enrichment for mega-loops and interchromosomal hubs that contain genes associated with neutrophil-specific enhancer repertoires and an inflammatory gene program. On the basis of these observations, we propose that in neutrophil progenitors, loop-extrusion programs produce lineage-specific chromatin architectures that permit the packing of chromosomes into geometrically confined lobular structures. Our data also provide a blueprint for the assembly of polymorphonuclear structures, and point to the possibility of engineering de novo nuclear shapes to facilitate the migration of effector cells in densely populated tumorigenic environments.

Loss of ATRX promotes aggressive features of osteosarcoma with increased NF-κB signaling and integrin binding.

Osteosarcoma (OS) is a lethal disease with few known targeted therapies. Here, we show that decreased ATRX expression is associated with more aggressive tumor cell phenotypes, including increased growth, migration, invasion, and metastasis. These phenotypic changes correspond with activation of NF-κB signaling, extracellular matrix remodeling, increased integrin αvß3 expression, and ETS family transcription factor binding. Here, we characterize these changes in vitro, in vivo, and in a data set of human OS patients. This increased aggression substantially sensitizes ATRX-deficient OS cells to integrin signaling inhibition. Thus, ATRX plays an important tumor-suppression role in OS, and loss of function of this gene may underlie new therapeutic vulnerabilities. The relationship between ATRX expression and integrin binding, NF-κB activation, and ETS family transcription factor binding has not been described in previous studies and may impact the pathophysiology of other diseases with ATRX loss, including other cancers and the ATR-X α thalassemia intellectual disability syndrome.

2-Hydroxyglutarate destabilizes chromatin regulatory landscape and lineage fidelity to promote cellular heterogeneity.

The epigenome delineates lineage-specific transcriptional programs and restricts cell plasticity to prevent non-physiological cell fate transitions. Although cell diversification fosters tumor evolution and therapy resistance, upstream mechanisms that regulate the stability and plasticity of the cancer epigenome remain elusive. Here we show that 2-hydroxyglutarate (2HG) not only suppresses DNA repair but also mediates the high-plasticity chromatin landscape. A combination of single-cell epigenomics and multi-omics approaches demonstrates that 2HG disarranges otherwise well-preserved stable nucleosome positioning and promotes cell-to-cell variability. 2HG induces loss of motif accessibility to the luminal-defining transcriptional factors FOXA1, FOXP1, and GATA3 and a shift from luminal to basal-like gene expression. Breast tumors with high 2HG exhibit enhanced heterogeneity with undifferentiated epigenomic signatures linked to adverse prognosis. Further, ascorbate-2-phosphate (A2P) eradicates heterogeneity and impairs growth of high 2HG-producing breast cancer cells. These findings suggest 2HG as a key determinant of cancer plasticity and provide a rational strategy to counteract tumor cell evolution.

Construction and Evaluation of Robust Interpretation Models for Breast Cancer Metastasis Prediction.

Interpretability of machine learning (ML) models represents the extent to which a model's decision-making process can be understood by model developers and/or end users. Transcriptomics-based cancer prognosis models, for example, while achieving good accuracy, are usually hard to interpret, due to the high-dimensional feature space and the complexity of models. As interpretability is critical for the transparency and fairness of ML models, several algorithms have been proposed to improve the interpretability of arbitrary classifiers. However, evaluation of these algorithms often requires substantial domain knowledge. Here, we propose a breast cancer metastasis prediction model using a very small number of biologically interpretable features, and a simple yet novel model interpretation approach that can provide personalized interpretations. In addition, we contributed, to the best of our knowledge, the first method to quantitatively compare different interpretation algorithms. Experimental results show that our model not only achieved competitive prediction accuracy, but also higher inter-classifier interpretation consistency than state-of-the-art interpretation methods. Importantly, our interpretation results can improve the generalizability of the prediction models. Overall, this work provides several novel ideas to construct and evaluate interpretable ML models that can be valuable to both the cancer machine learning community and related application domains.

Gene selection for optimal prediction of cell position in tissues from single-cell transcriptomics data.

Single-cell RNA-sequencing (scRNAseq) technologies are rapidly evolving. Although very informative, in standard scRNAseq experiments, the spatial organization of the cells in the tissue of origin is lost. Conversely, spatial RNA-seq technologies designed to maintain cell localization have limited throughput and gene coverage. Mapping scRNAseq to genes with spatial information increases coverage while providing spatial location. However, methods to perform such mapping have not yet been benchmarked. To fill this gap, we organized the DREAM Single-Cell Transcriptomics challenge focused on the spatial reconstruction of cells from the Drosophila embryo from scRNAseq data, leveraging as silver standard, genes with in situ hybridization data from the Berkeley Drosophila Transcription Network Project reference atlas. The 34 participating teams used diverse algorithms for gene selection and location prediction, while being able to correctly localize clusters of cells. Selection of predictor genes was essential for this task. Predictor genes showed a relatively high expression entropy, high spatial clustering and included prominent developmental genes such as gap and pair-rule genes and tissue markers. Application of the top 10 methods to a zebra fish embryo dataset yielded similar performance and statistical properties of the selected genes than in the Drosophila data. This suggests that methods developed in this challenge are able to extract generalizable properties of genes that are useful to accurately reconstruct the spatial arrangement of cells in tissues.

Network-Based Single-Cell RNA-Seq Data Imputation Enhances Cell Type Identification.

Single-cell RNA sequencing is a powerful technology for obtaining transcriptomes at single-cell resolutions. However, it suffers from dropout events (i.e., excess zero counts) since only a small fraction of transcripts get sequenced in each cell during the sequencing process. This inherent sparsity of expression profiles hinders further characterizations at cell/gene-level such as cell type identification and downstream analysis. To alleviate this dropout issue we introduce a network-based method, netImpute, by leveraging the hidden information in gene co-expression networks to recover real signals. netImpute employs Random Walk with Restart (RWR) to adjust the gene expression level in a given cell by borrowing information from its neighbors in a gene co-expression network. Performance evaluation and comparison with existing tools on simulated data and seven real datasets show that netImpute substantially enhances clustering accuracy and data visualization clarity, thanks to its effective treatment of dropouts. While the idea of netImpute is general and can be applied with other types of networks such as cell co-expression network or protein-protein interaction (PPI) network, evaluation results show that gene co-expression network is consistently more beneficial, presumably because PPI network usually lacks cell type context, while cell co-expression network can cause information loss for rare cell types. Evaluation results on several biological datasets show that netImpute can more effectively recover missing transcripts in scRNA-seq data and enhance the identification and visualization of heterogeneous cell types than existing methods.

Spatial mapping of single cells in the Drosophila embryo from transcriptomic data based on topological consistency.

The advancement in single-cell RNA sequencing technologies allow us to obtain transcriptome at single cell resolution. However, the original spatial context of cells, a crucial knowledge for understanding cellular and tissue-level functions, is often lost during sequencing. To address this issue, the DREAM Single Cell Transcriptomics Challenge launched a community-wide effort to seek computational solutions for spatial mapping of single cells in tissues using single-cell RNAseq (scRNA-seq) data and a reference atlas obtained from in situ hybridization data. As a top-performing team in this competition, we approach this problem in three steps. The first step involves identifying a set of most informative genes based on the consistency between gene expression similarity and cell proximity. For this step, we propose two different approaches, i.e., an unsupervised approach that does not utilize the gold standard location of the cells provided by the challenge organizers, and a supervised approach that relies on the gold standard locations. In the second step, a Particle Swarm Optimization algorithm is used to optimize the weights of different genes in order to maximize matches between the predicted locations and the gold standard locations. Finally, the information embedded in the cell topology is used to improve the predicted cell-location scores by weighted averaging of scores from neighboring locations. Evaluation results based on DREAM scores show that our method accurately predicts the location of single cells, and the predictions lead to successful recovery of the spatial expression patterns for most of landmark genes. In addition, investigating the selected genes demonstrates that most predictive genes are cluster specific, and stable across our supervised and unsupervised gene selection frameworks. Overall, the promising results obtained by our methods in DREAM challenge demonstrated that topological consistency is a useful concept in identifying marker genes and constructing predictive models for spatial mapping of single cells.