Explore the vast range of titles available.

Background Deep learning has emerged as a versatile approach for predicting complex biological phenomena. However, its utility for biological discovery has so far been limited, given that generic deep neural networks provide little insight into the biological mechanisms that underlie a successful prediction. Here we demonstrate deep learning on biological networks, where every node has a molecular equivalent, such as a protein or gene, and every edge has a mechanistic interpretation, such as a regulatory interaction along a signaling pathway. Results With knowledge-primed neural networks (KPNNs), we exploit the ability of deep learning algorithms to assign meaningful weights in multi-layered networks, resulting in a widely applicable approach for interpretable deep learning. We present a learning method that enhances the interpretability of trained KPNNs by stabilizing node weights in the presence of redundancy, enhancing the quantitative interpretability of node weights, and controlling for uneven connectivity in biological networks. We validate KPNNs on simulated data with known ground truth and demonstrate their practical use and utility in five biological applications with single-cell RNA-seq data for cancer and immune cells. Conclusions We introduce KPNNs as a method that combines the predictive power of deep learning with the interpretability of biological networks. While demonstrated here on single-cell sequencing data, this method is broadly relevant to other research areas where prior domain knowledge can be represented as networks.

DNA methylation is a widely investigated epigenetic mark with important roles in development and disease. High-throughput assays enable genome-scale DNA methylation analysis in large numbers of samples. Here, we describe a new version of our RnBeads software - an R/Bioconductor package that implements start-to-finish analysis workflows for Infinium microarrays and various types of bisulfite sequencing. RnBeads 2.0 ( https://rnbeads.org/ ) provides additional data types and analysis methods, new functionality for interpreting DNA methylation differences, improved usability with a novel graphical user interface, and better use of computational resources. We demonstrate RnBeads 2.0 in four re-runnable use cases focusing on cell differentiation and cancer.

ChIPmentation combines chromatin immunoprecipitation with on-bead tagmentation for rapid and highly robust ChIP-seq library preparation. Chromatin immunoprecipitation followed by sequencing (ChIP-seq) is widely used to map histone marks and transcription factor binding throughout the genome. Here we present ChIPmentation, a method that combines chromatin immunoprecipitation with sequencing library preparation by Tn5 transposase ('tagmentation'). ChIPmentation introduces sequencing-compatible adaptors in a single-step reaction directly on bead-bound chromatin, which reduces time, cost and input requirements, thus providing a convenient and broadly useful alternative to existing ChIP-seq protocols.

Tissue-resident macrophages support embryonic development and tissue homeostasis and repair. The mechanisms that control their differentiation remain unclear. We report here that erythro-myeloid progenitors in mice generate premacrophages (pMacs) that simultaneously colonize the whole embryo from embryonic day 9.5 in a chemokine-receptor–dependent manner. The core macrophage program initiated in pMacs is rapidly diversified as expression of transcriptional regulators becomes tissue-specific in early macrophages. This process appears essential for macrophage specification and maintenance, as inactivation of Id3 impairs the development of liver macrophages and results in selective Kupffer cell deficiency in adults. We propose that macrophage differentiation is an integral part of organogenesis, as colonization of organ anlagen by pMacs is followed by their specification into tissue macrophages, hereby generating the macrophage diversity observed in postnatal tissues.

Targeted protein degradation is a new therapeutic modality based on drugs that destabilize proteins by inducing their proximity to E3 ubiquitin ligases. Of particular interest are molecular glues that can degrade otherwise unligandable proteins by orchestrating direct interactions between target and ligase. However, their discovery has so far been serendipitous, thus hampering broad translational efforts. Here, we describe a scalable strategy toward glue degrader discovery that is based on chemical screening in hyponeddylated cells coupled to a multi-omics target deconvolution campaign. This approach led us to identify compounds that induce ubiquitination and degradation of cyclin K by prompting an interaction of CDK12–cyclin K with a CRL4B ligase complex. Notably, this interaction is independent of a dedicated substrate receptor, thus functionally segregating this mechanism from all described degraders. Collectively, our data outline a versatile and broadly applicable strategy to identify degraders with nonobvious mechanisms and thus empower future drug discovery efforts. Chemical profiling in hyponeddylated cells coupled with multi-omics target deconvolution led to the identification of molecular glue degraders of cyclin K that function by inducing proximity between the CRL adaptor DDB1 and a CDK12–cyclin K complex.

Key Points Recent technological advances make it possible to map DNA methylation in essentially any cell type, tissue or organism. Computational methods and software tools are essential for processing, analysing and interpreting large-scale DNA methylation data sets. Tailored software tools are now available for processing data obtained with all common methods for genome-wide DNA methylation mapping (including bisulphite sequencing and the Infinium assay). Bioinformatic methods for visualization of DNA methylation data facilitate quality assessment and help to pinpoint global trends in the data. By combining stringent statistical methods with computational and experimental validation, researchers can establish accurate lists of differentially methylated regions for a phenotype of interest. Biological interpretation of differential DNA methylation is aided by computational tools for data exploration and enrichment analysis. Large community projects are currently generating reference epigenome maps for many different cell types; the interpretation of these maps will require a comprehensive effort in functional epigenomics. The analysis and interpretation of genome-wide DNA methylation data poses unique bioinformatics challenges. In this article, the tools that are available for processing, visualizing and interpreting these epigenetic data sets are discussed, and the relative advantages of various methods are considered. DNA methylation is an epigenetic mark that has suspected regulatory roles in a broad range of biological processes and diseases. The technology is now available for studying DNA methylation genome-wide, at a high resolution and in a large number of samples. This Review discusses relevant concepts, computational methods and software tools for analysing and interpreting DNA methylation data. It focuses not only on the bioinformatic challenges of large epigenome-mapping projects and epigenome-wide association studies but also highlights software tools that make genome-wide DNA methylation mapping more accessible for laboratories with limited bioinformatics experience.

Tumor associated inflammation predicts response to immune checkpoint blockade in human melanoma. Current theories on regulation of inflammation center on anti-tumor T cell responses. Here we show that tumor associated B cells are vital to melanoma associated inflammation. Human B cells express pro- and anti-inflammatory factors and differentiate into plasmablast-like cells when exposed to autologous melanoma secretomes in vitro. This plasmablast-like phenotype can be reconciled in human melanomas where plasmablast-like cells also express T cell-recruiting chemokines CCL3, CCL4, CCL5. Depletion of B cells in melanoma patients by anti-CD20 immunotherapy decreases tumor associated inflammation and CD8 + T cell numbers. Plasmablast-like cells also increase PD-1 + T cell activation through anti-PD-1 blockade in vitro and their frequency in pretherapy melanomas predicts response and survival to immune checkpoint blockade. Tumor associated B cells therefore orchestrate and sustain melanoma inflammation and may represent a predictor for survival and response to immune checkpoint blockade therapy. The regulation of tumor inflammation is incompletely understood and the role of B cells is unclear. Here, the authors show that a specific subtype of B cells is induced in melanoma and required to recruit T lymphocytes and elicit inflammation.

Neural crest cells develop into tissues ranging from craniofacial bones to peripheral neurons. Combining single-cell RNA sequencing with spatial transcriptomics, Soldatov et al. analyzed how neural crest cells in mouse embryos decide among the various fates available to them (see the Perspective by Mayor). These multipotent cells become biased toward a given fate early on and step through a progression of binary decisions as their fate is refined. Competing fate programs coexist until increased synchronization favors one and repression disfavors the other. Science , this issue p. eaas9536 ; see also p. 937 Differentiation of cells of the neural crest proceeds through binary decisions that restrict and refine developmental paths. Neural crest cells are embryonic progenitors that generate numerous cell types in vertebrates. With single-cell analysis, we show that mouse trunk neural crest cells become biased toward neuronal lineages when they delaminate from the neural tube, whereas cranial neural crest cells acquire ectomesenchyme potential dependent on activation of the transcription factor Twist1. The choices that neural crest cells make to become sensory, glial, autonomic, or mesenchymal cells can be formalized as a series of sequential binary decisions. Each branch of the decision tree involves initial coactivation of bipotential properties followed by gradual shifts toward commitment. Competing fate programs are coactivated before cells acquire fate-specific phenotypic traits. Determination of a specific fate is achieved by increased synchronization of relevant programs and concurrent repression of competing fate programs.

Genome-wide mapping of 5-methylcytosine is of broad interest to many fields of biology and medicine. A variety of methods have been developed, and several have recently been advanced to genome-wide scale using arrays and next-generation sequencing approaches. We have previously reported reduced representation bisulfite sequencing (RRBS), a bisulfite-based protocol that enriches CG-rich parts of the genome, thereby reducing the amount of sequencing required while capturing the majority of promoters and other relevant genomic regions. The approach provides single-nucleotide resolution, is highly sensitive and provides quantitative DNA methylation measurements. This protocol should enable any standard molecular biology laboratory to generate RRBS libraries of high quality. Briefly, purified genomic DNA is digested by the methylation-insensitive restriction enzyme MspI to generate short fragments that contain CpG dinucleotides at the ends. After end-repair, A-tailing and ligation to methylated Illumina adapters, the CpG-rich DNA fragments (40–220 bp) are size selected, subjected to bisulfite conversion, PCR amplified and end sequenced on an Illumina Genome Analyzer. Note that alignment and analysis of RRBS sequencing reads are not covered in this protocol. The extremely low input requirements (10–300 ng), the applicability of the protocol to formalin-fixed and paraffin-embedded samples, and the technique's single-nucleotide resolution extends RRBS to a wide range of biological and clinical samples and research applications. The entire process of RRBS library construction takes ∼9 d.

The authors identified calreticulin mutations in the majority of patients with essential thrombocythemia and myelofibrosis who did not have JAK2 mutations. The mutation alters calreticulin protein, and cells expressing the mutant protein are more responsive to growth factors. Philadelphia chromosome–negative myeloproliferative neoplasms include polycythemia vera, essential thrombocythemia, and primary myelofibrosis. 1 A unique gain-of-function mutation in the Janus kinase 2 gene ( JAK2 ) is found in about three quarters of patients in whom these disease entities have been diagnosed. 2 , 3 The valine-to-phenylalanine (V617F) alteration constitutively activates JAK2, resulting in increased phosphorylation of its substrates and leading to increased cytokine responsiveness of myeloid cells. The JAK2 V617F mutation is present in approximately 95% of patients with polycythemia vera and in 50 to 60% of those with essential thrombocythemia or primary myelofibrosis. 4 In addition, somatic mutations of JAK2 exon 12 . . .