Decoding gene regulation in the fly brain

Evaluating methods for the prediction of cell-type-specific enhancers in the mammalian cortex

Identifying cell-type-specific enhancers is critical for developing genetic tools to study the mammalian brain. We organized the "Brain Initiative Cell Census Network (BICCN) Challenge: Predicting Functional Cell Type-Specific Enhancers from Cross-Species Multi-Omics" to evaluate machine learning and feature-based methods for nominating enhancer sequences targeting mouse cortical cell types. Methods were assessed using in vivo data from hundreds of adeno-associated virus (AAV)-packaged, retro-orbitally delivered enhancers. Open chromatin was the strongest predictor of functional enhancers, while sequence models improved prediction of non-functional enhancers and identified cell-type-specific transcription factor codes to inform in silico enhancer design. This challenge establishes a benchmark for enhancer prioritization and highlights computational and molecular features critical for identifying functional cortical enhancers, advancing efforts to map and manipulate gene regulation in the mammalian cortex.

Predicting gene expression from DNA sequence using deep learning models

Transcription of genes is regulated by DNA elements such as promoters and enhancers, the activity of which are in turn controlled by many transcription factors. Owing to the highly complex combinatorial logic involved, it has been difficult to construct computational models that predict gene activity from DNA sequence. Recent advances in deep learning techniques applied to data from epigenome mapping and high-throughput reporter assays have made substantial progress towards addressing this complexity. Such models can capture the regulatory grammar with remarkable accuracy and show great promise in predicting the effects of non-coding variants, uncovering detailed molecular mechanisms of gene regulation and designing synthetic regulatory elements for biotechnology. Here, we discuss the principles of these approaches, the types of training data sets that are available and the strengths and limitations of different approaches.

Advances and applications in single-cell and spatial genomics

The applications of single-cell and spatial technologies in recent times have revolutionized the present understanding of cellular states and the cellular heterogeneity inherent in complex biological systems. These advancements offer unprecedented resolution in the examination of the functional genomics of individual cells and their spatial context within tissues. In this review, we have comprehensively discussed the historical development and recent progress in the field of single-cell and spatial genomics. We have reviewed the breakthroughs in single-cell multi-omics technologies, spatial genomics methods, and the computational strategies employed toward the analyses of single-cell atlas data. Furthermore, we have highlighted the advances made in constructing cellular atlases and their clinical applications, particularly in the context of disease. Finally, we have discussed the emerging trends, challenges, and opportunities in this rapidly evolving field.

Differential neuronal vulnerability to C9orf72 repeat expansion driven by Xbp1-induced endoplasmic reticulum-associated degradation

Neurodegenerative diseases are characterized by the localized loss of neurons. Why cell death is triggered only in specific neuronal populations and whether it is the response to toxic insults or the initial cellular state that determines their vulnerability is unknown. To understand individual cell responses to disease, we profiled their transcriptional signatures throughout disease development in a Drosophila model of C9orf72 (G4C2) repeat expansion (C9), the most common genetic cause of frontotemporal dementia and amyotrophic lateral sclerosis. We identified neuronal populations specifically vulnerable or resistant to C9 expression and found an upregulation of protein homeostasis pathways in resistant neurons at baseline. Overexpression of Xbp1s, a key regulator of the unfolded protein response and a central node in the resistance network, rescues C9 toxicity. This study shows that neuronal vulnerability depends on the intrinsic transcriptional state of neurons and that leveraging resistant neurons' properties can boost resistance in vulnerable neurons.

Interpreting regulatory mechanisms of Hippo signaling through a deep learning sequence model

Signaling pathway components are well studied, but how they mediate cell-type-specific transcription responses is an unresolved problem. Using the Hippo pathway in mouse trophoblast stem cells as a model, we show that the DNA binding of signaling effectors is driven by cell-type-specific sequence rules that can be learned genome wide by deep learning models. Through model interpretation and experimental validation, we show that motifs for the cell-type-specific transcription factor TFAP2C enhance TEAD4/YAP1 binding in a nucleosome-range and distance-dependent manner, driving synergistic enhancer activation. We also discovered that Tead double motifs are widespread, highly active canonical response elements. Molecular dynamics simulations suggest that TEAD4 binds them cooperatively through surprisingly labile protein-protein interactions that depend on the DNA template. These results show that the response to signaling pathways is encoded in the cis-regulatory sequences and that interpreting the rules reveals insights into the mechanisms by which signaling effectors influence cell-type-specific enhancer activity.

Synaptic deregulation of cholinergic projection neurons causes olfactory dysfunction across five fly Parkinsonism models

The classical diagnosis of Parkinsonism is based on motor symptoms that are the consequence of nigrostriatal pathway dysfunction and reduced dopaminergic output. However, a decade prior to the emergence of motor issues, patients frequently experience non-motor symptoms, such as a reduced sense of smell (hyposmia). The cellular and molecular bases for these early defects remain enigmatic. To explore this, we developed a new collection of five fruit fly models of familial Parkinsonism and conducted single-cell RNA sequencing on young brains of these models. Interestingly, cholinergic projection neurons are the most vulnerable cells, and genes associated with presynaptic function are the most deregulated. Additional single nucleus sequencing of three specific brain regions of Parkinson's disease patients confirms these findings. Indeed, the disturbances lead to early synaptic dysfunction, notably affecting cholinergic olfactory projection neurons crucial for olfactory function in flies. Correcting these defects specifically in olfactory cholinergic interneurons in flies or inducing cholinergic signaling in Parkinson mutant human induced dopaminergic neurons in vitro using nicotine, both rescue age-dependent dopaminergic neuron decline. Hence, our research uncovers that one of the earliest indicators of disease in five different models of familial Parkinsonism is synaptic dysfunction in higher-order cholinergic projection neurons and this contributes to the development of hyposmia. Furthermore, the shared pathways of synaptic failure in these cholinergic neurons ultimately contribute to dopaminergic dysfunction later in life.

RNAi-based screen for pigmentation in Drosophila melanogaster reveals regulators of brain dopamine and sleep

The dopaminergic system has been extensively studied for its role in behavior in animals as well as human neuropsychiatric and neurological diseases. However, we still know little about how dopamine levels are tightly regulated in vivo . To identify novel regulators of dopamine, we utilized Drosophila melanogaster cuticle pigmentation as a readout, where dopamine is a precursor to melanin. We measured dopamine from genes known to be critical for cuticle pigmentation and performed an RNAi-based screen to identify new regulators of pigmentation. We found 153 potential pigmentation genes, which were enriched for conserved homologs and disease- associated genes as well as developmental signaling pathways and mitochondria-associated proteins. From 35 prioritized candidates, we found 10 caused significant reduction in head dopamine levels and one caused an increase. Two genes, clueless and mask (multiple ankyrin repeats single KH domain), upon knockdown, reduced dopamine levels in the brain. Further examination suggests that Mask regulates the transcription of the rate-limiting dopamine synthesis enzyme, tyrosine hydroxylase , and its knockdown causes dopamine-dependent sleep phenotypes. In summary, by studying genes that affect cuticle pigmentation, a phenotype seemingly unrelated to the nervous system, we were able to identify several genes that affect dopamine metabolism as well as a novel regulator of behavior.

Evaluating Methods for the Prediction of Cell Type-Specific Enhancers in the Mammalian Cortex

Identifying cell type-specific enhancers in the brain is critical to building genetic tools for investigating the mammalian brain. Computational methods for functional enhancer prediction have been proposed and validated in the fruit fly and not yet the mammalian brain. We organized the 'Brain Initiative Cell Census Network (BICCN) Challenge: Predicting Functional Cell Type-Specific Enhancers from Cross-Species Multi-Omics' to assess machine learning and feature-based methods designed to nominate enhancer DNA sequences to target cell types in the mouse cortex. Methods were evaluated based on in vivo validation data from hundreds of cortical cell type-specific enhancers that were previously packaged into individual AAV vectors and retro-orbitally injected into mice. We find that open chromatin was a key predictor of functional enhancers, and sequence models improved prediction of non-functional enhancers that can be deprioritized as opposed to pursued for in vivo testing. Sequence models also identified cell type-specific transcription factor codes that can guide designs of in silico enhancers. This community challenge establishes a benchmark for enhancer prioritization algorithms and reveals computational approaches and molecular information that are crucial for identifying functional enhancers in mammalian cortical cell types. The results of this challenge bring us closer to understanding the complex gene regulatory landscape of the mammalian cortex and to designing more efficient genetic tools to target cortical cell types.

Design principles of gene circuits for longevity

Aging is a dynamic process that is driven by cellular damage and disruption of homeostatic gene regulatory networks (GRNs). Traditional studies often focus on individual genes, but understanding their interplay is key to unraveling the mechanisms of aging. This review explores the gene circuits that influence longevity and highlights the role of feedback loops in maintaining cellular balance. The SIR2-HAP circuit in yeast serves as a model to explore how mutual inhibition between pathways influences aging trajectories and how engineering stable fixed points or oscillations within these circuits can extend lifespan. Feedback loops crucial for maintaining homeostasis are also reviewed, and we highlight how their destabilization accelerates aging. By leveraging systems and synthetic biology, strategies are proposed that may stabilize these loops within single cells, thereby enhancing their resilience to aging-related damage.

Transfer learning reveals sequence determinants of the quantitative response to transcription factor dosage

Deep learning models have advanced our ability to predict cell-type-specific chromatin patterns from transcription factor (TF) binding motifs, but their application to perturbed contexts remains limited. We applied transfer learning to predict how concentrations of the dosage-sensitive TFs TWIST1 and SOX9 affect regulatory element (RE) chromatin accessibility in facial progenitor cells, achieving near-experimental accuracy. High-affinity motifs that allow for heterotypic TF co-binding and are concentrated at the center of REs buffer against quantitative changes in TF dosage and predict unperturbed accessibility. Conversely, low-affinity or homotypic binding motifs distributed throughout REs drive sensitive responses with minimal impact on unperturbed accessibility. Both buffering and sensitizing features display purifying selection signatures. We validated these sequence features through reporter assays and demonstrated that TF-nucleosome competition can explain low-affinity motifs' sensitizing effects. This combination of transfer learning and quantitative chromatin response measurements provides a novel approach for uncovering additional layers of the cis-regulatory code.

Molecular profiling of invertebrate glia

Caenorhabditis elegans and Drosophila melanogaster are powerful experimental models for uncovering fundamental tenets of nervous system organization and function. Findings over the last two decades show that molecular and cellular features are broadly conserved between invertebrates and vertebrates, indicating that insights derived from invertebrate models can broadly inform our understanding of glial operating principles across diverse species. In recent years, these model systems have led to exciting discoveries in glial biology and mechanisms of glia-neuron interactions. Here, we summarize studies that have applied current state-of-the-art "-omics" techniques to C. elegans and D. melanogaster glia. Coupled with the remarkable acceleration in the pace of mechanistic studies of glia biology in recent years, these indicate that invertebrate glia also exhibit striking molecular complexity, specificity, and heterogeneity. We provide an overview of these studies and discuss their implications as well as emerging questions where C. elegans and D. melanogaster are well-poised to fill critical knowledge gaps in our understanding of glial biology.

Imp/IGF2BP and Syp/SYNCRIP temporal RNA interactomes uncover combinatorial networks of regulators of Drosophila brain development

Temporal patterning of neural progenitors is an evolutionarily conserved mechanism generating neural diversity. In Drosophila, postembryonic neurogenesis requires the RNA binding proteins (RBPs) Imp/IGF2BP and Syp/SYNCRIP. However, how they coachieve their function is not well understood. Here, we elucidate the in vivo temporal RNA interactome landscapes of Imp and Syp during larval brain development. Imp and Syp bind a highly overlapping set of conserved mRNAs encoding proteins involved in neurodevelopment. We identify transcripts differentially occupied by Imp/Syp over time, featuring a network of known and previously unknown candidate temporal regulators that are post-transcriptionally regulated by Imp/Syp. Furthermore, the physical and coevolutionary relationships between Imp and Syp binding sites reveal a combinatorial, rather than competitive, mode of molecular interplay. Our study establishes an in vivo framework for dissecting the temporal coregulation of RBP networks as well as providing a resource for understanding neural fate specification.

Enhancer-driven cell type comparison reveals similarities between the mammalian and bird pallium

Combinations of transcription factors govern the identity of cell types, which is reflected by genomic enhancer codes. We used deep learning to characterize these enhancer codes and devised three metrics to compare cell types in the telencephalon across amniotes. To this end, we generated single-cell multiome and spatially resolved transcriptomics data of the chicken telencephalon. Enhancer codes of orthologous nonneuronal and γ-aminobutyric acid-mediated (GABAergic) cell types show a high degree of similarity across amniotes, whereas excitatory neurons of the mammalian neocortex and avian pallium exhibit varying degrees of similarity. Enhancer codes of avian mesopallial neurons are most similar to those of mammalian deep-layer neurons. With this study, we present generally applicable deep learning approaches to characterize and compare cell types on the basis of genomic regulatory sequences.

Isoforms of terminal selector Mamo control axon guidance during adult Drosophila memory center construction via Semaphorin-1a

In animals undergoing metamorphosis, the appearance of the nervous system is coincidently transformed by the morphogenesis of neurons. Such morphogenic alterations are exemplified in three types of intrinsic neurons in the Drosophila memory center. In contrast to the well-characterized remodeling of γ neurons, the morphogenesis of α/β and α'/β' neurons has not been adequately explored. Here, we show that mamo, a BTB-zinc finger transcription factor that acts as a terminal selector for α'/β' neurons, controls the formation of the correct axonal pattern of α'/β' neurons. Intriguingly, specific Mamo isoforms are preferentially expressed in α'/β' neurons to regulate the expression of axon guidance molecule Semaphorin-1a. This action directs proper axon guidance in α'/β' neurons, which is also crucial for wiring of α'/β' neurons with downstream neurons. Taken together, our results provide molecular insights into how neurons establish correct axonal patterns in circuitry assembly during adult memory center construction.

Synergistic activation by Glass and Pointed promotes neuronal identity in the Drosophila eye disc

The integration of extrinsic signaling with cell-intrinsic transcription factors can direct progenitor cells to differentiate into distinct cell fates. In the developing Drosophila eye, differentiation of photoreceptors R1-R7 requires EGFR signaling mediated by the transcription factor Pointed, and our single-cell RNA-Seq analysis shows that the same photoreceptors require the eye-specific transcription factor Glass. We find that ectopic expression of Glass and activation of EGFR signaling synergistically induce neuronal gene expression in the wing disc in a Pointed-dependent manner. Targeted DamID reveals that Glass and Pointed share many binding sites in the genome of developing photoreceptors. Comparison with transcriptomic data shows that Pointed and Glass induce photoreceptor differentiation through intermediate transcription factors, including the redundant homologs Scratch and Scrape, as well as directly activating neuronal effector genes. Our data reveal synergistic activation of a multi-layered transcriptional network as the mechanism by which EGFR signaling induces neuronal identity in Glass-expressing cells.

Enhancer-driven gene regulatory networks inference from single-cell RNA-seq and ATAC-seq data

Deciphering the intricate relationships between transcription factors (TFs), enhancers, and genes through the inference of enhancer-driven gene regulatory networks (eGRNs) is crucial in understanding gene regulatory programs in a complex biological system. This study introduces STREAM, a novel method that leverages a Steiner forest problem model, a hybrid biclustering pipeline, and submodular optimization to infer eGRNs from jointly profiled single-cell transcriptome and chromatin accessibility data. Compared to existing methods, STREAM demonstrates enhanced performance in terms of TF recovery, TF-enhancer linkage prediction, and enhancer-gene relation discovery. Application of STREAM to an Alzheimer's disease dataset and a diffuse small lymphocytic lymphoma dataset reveals its ability to identify TF-enhancer-gene relations associated with pseudotime, as well as key TF-enhancer-gene relations and TF cooperation underlying tumor cells.

Epigenetic alterations affecting hematopoietic regulatory networks as drivers of mixed myeloid/lymphoid leukemia

Leukemias with ambiguous lineage comprise several loosely defined entities, often without a clear mechanistic basis. Here, we extensively profile the epigenome and transcriptome of a subgroup of such leukemias with CpG Island Methylator Phenotype. These leukemias exhibit comparable hybrid myeloid/lymphoid epigenetic landscapes, yet heterogeneous genetic alterations, suggesting they are defined by their shared epigenetic profile rather than common genetic lesions. Gene expression enrichment reveals similarity with early T-cell precursor acute lymphoblastic leukemia and a lymphoid progenitor cell of origin. In line with this, integration of differential DNA methylation and gene expression shows widespread silencing of myeloid transcription factors. Moreover, binding sites for hematopoietic transcription factors, including CEBPA, SPI1 and LEF1, are uniquely inaccessible in these leukemias. Hypermethylation also results in loss of CTCF binding, accompanied by changes in chromatin interactions involving key transcription factors. In conclusion, epigenetic dysregulation, and not genetic lesions, explains the mixed phenotype of this group of leukemias with ambiguous lineage. The data collected here constitute a useful and comprehensive epigenomic reference for subsequent studies of acute myeloid leukemias, T-cell acute lymphoblastic leukemias and mixed-phenotype leukemias.

Big data and deep learning for RNA biology

The exponential growth of big data in RNA biology (RB) has led to the development of deep learning (DL) models that have driven crucial discoveries. As constantly evidenced by DL studies in other fields, the successful implementation of DL in RB depends heavily on the effective utilization of large-scale datasets from public databases. In achieving this goal, data encoding methods, learning algorithms, and techniques that align well with biological domain knowledge have played pivotal roles. In this review, we provide guiding principles for applying these DL concepts to various problems in RB by demonstrating successful examples and associated methodologies. We also discuss the remaining challenges in developing DL models for RB and suggest strategies to overcome these challenges. Overall, this review aims to illuminate the compelling potential of DL for RB and ways to apply this powerful technology to investigate the intriguing biology of RNA more effectively.

Single-cell transcriptomics dissecting the development and evolution of nervous system in insects

Insects can display a vast repertoire of complex and adaptive behaviors crucial for survival and reproduction. Yet, how the neural circuits underlying insect behaviors are assembled throughout development and remodeled during evolution remains largely obscure. The advent of single-cell transcriptomics has opened new paths to illuminate these historically intractable questions. Insect behavior is governed by its brain, whose functional complexity is realized through operations across multiple levels, from the molecular and cellular to the circuit and organ. Single-cell transcriptomics enables dissecting brain functions across all these levels and allows tracking regulatory dynamics throughout development and under perturbation. In this review, we mainly focus on the achievements of single-cell transcriptomics in dissecting the molecular and cellular architectures of nervous systems in representative insects, then discuss its applications in tracking the developmental trajectory and functional evolution of insect brains.

Chromatin accessibility during human first-trimester neurodevelopment

The human brain develops through a tightly organized cascade of patterning events, induced by transcription factor expression and changes in chromatin accessibility. Although gene expression across the developing brain has been described at single-cell resolution1, similar atlases of chromatin accessibility have been primarily focused on the forebrain2-4. Here we describe chromatin accessibility and paired gene expression across the entire developing human brain during the first trimester (6-13 weeks after conception). We defined 135 clusters and used multiomic measurements to link candidate cis-regulatory elements to gene expression. The number of accessible regions increased both with age and along neuronal differentiation. Using a convolutional neural network, we identified putative functional transcription factor-binding sites in enhancers characterizing neuronal subtypes. We applied this model to cis-regulatory elements linked to ESRRB to elucidate its activation mechanism in the Purkinje cell lineage. Finally, by linking disease-associated single nucleotide polymorphisms to cis-regulatory elements, we validated putative pathogenic mechanisms in several diseases and identified midbrain-derived GABAergic neurons as being the most vulnerable to major depressive disorder-related mutations. Our findings provide a more detailed view of key gene regulatory mechanisms underlying the emergence of brain cell types during the first trimester and a comprehensive reference for future studies related to human neurodevelopment.

Clock-dependent chromatin accessibility rhythms regulate circadian transcription

Chromatin organization plays a crucial role in gene regulation by controlling the accessibility of DNA to transcription machinery. While significant progress has been made in understanding the regulatory role of clock proteins in circadian rhythms, how chromatin organization affects circadian rhythms remains poorly understood. Here, we employed ATAC-seq (Assay for Transposase-Accessible Chromatin with Sequencing) on FAC-sorted Drosophila clock neurons to assess genome-wide chromatin accessibility at dawn and dusk over the circadian cycle. We observed significant oscillations in chromatin accessibility at promoter and enhancer regions of hundreds of genes, with enhanced accessibility either at dusk or dawn, which correlated with their peak transcriptional activity. Notably, genes with enhanced accessibility at dusk were enriched with E-box motifs, while those more accessible at dawn were enriched with VRI/PDP1-box motifs, indicating that they are regulated by the core circadian feedback loops, PER/CLK and VRI/PDP1, respectively. Further, we observed a complete loss of chromatin accessibility rhythms in per01 null mutants, with chromatin consistently accessible at both dawn and dusk, underscoring the critical role of Period protein in driving chromatin compaction during the repression phase at dawn. Together, this study demonstrates the significant role of chromatin organization in circadian regulation, revealing how the interplay between clock proteins and chromatin structure orchestrates the precise timing of biological processes throughout the day. This work further implies that variations in chromatin accessibility might play a central role in the generation of diverse circadian gene expression patterns in clock neurons.

Iterative assay for transposase-accessible chromatin by sequencing to isolate functionally relevant neuronal subtypes

The Drosophila brain contains tens of thousands of distinct cell types. Thousands of different transgenic lines reproducibly target specific neuron subsets, yet most still express in several cell types. Furthermore, most lines were developed without a priori knowledge of where the transgenes would be expressed. To aid in the development of cell type-specific tools for neuronal identification and manipulation, we developed an iterative assay for transposase-accessible chromatin (ATAC) approach. Open chromatin regions (OCRs) enriched in neurons, compared to whole bodies, drove transgene expression preferentially in subsets of neurons. A second round of ATAC-seq from these specific neuron subsets revealed additional enriched OCR2s that further restricted transgene expression within the chosen neuron subset. This approach allows for continued refinement of transgene expression, and we used it to identify neurons relevant for sleep behavior. Furthermore, this approach is widely applicable to other cell types and to other organisms.

Construction of single-cell cross-species chromatin accessibility landscapes with combinatorial-hybridization-based ATAC-seq

Despite recent advances in single-cell genomics, the lack of maps for single-cell candidate cis-regulatory elements (cCREs) in non-mammal species has limited our exploration of conserved regulatory programs across vertebrates and invertebrates. Here, we developed a combinatorial-hybridization-based method for single-cell assay for transposase-accessible chromatin using sequencing (scATAC-seq) named CH-ATAC-seq, enabling the construction of single-cell accessible chromatin landscapes for zebrafish, Drosophila, and earthworms (Eisenia andrei). By integrating scATAC censuses of humans, monkeys, and mice, we systematically identified 152 distinct main cell types and around 0.8 million cell-type-specific cCREs. Our analysis provided insights into the conservation of neural, muscle, and immune lineages across species, while epithelial cells exhibited a higher organ-origin heterogeneity. Additionally, a large-scale gene regulatory network (GRN) was constructed in four vertebrates by integrating scRNA-seq censuses. Overall, our study provides a valuable resource for comparative epigenomics, identifying the evolutionary conservation and divergence of gene regulation across different species.

Stochastic neuro-fuzzy system implemented in memristor crossbar arrays

Neuro-symbolic artificial intelligence has garnered considerable attention amid increasing industry demands for high-performance neural networks that are interpretable and adaptable to previously unknown problem domains with minimal reconfiguration. However, implementing neuro-symbolic hardware is challenging due to the complexity in symbolic knowledge representation and calculation. We experimentally demonstrated a memristor-based neuro-fuzzy hardware based on TiN/TaOx/HfOx/TiN chips that is superior to its silicon-based counterpart in terms of throughput and energy efficiency by using array topological structure for knowledge representation and physical laws for computing. Intrinsic memristor variability is fully exploited to increase robustness in knowledge representation. A hybrid in situ training strategy is proposed for error minimizing in training. The hardware adapts easier to a previously unknown environment, achieving ~6.6 times faster convergence and ~6 times lower error than deep learning. The hardware energy efficiency is over two orders of magnitude greater than field-programmable gate arrays. This research greatly extends the capability of memristor-based neuromorphic computing systems in artificial intelligence.

Machine Learning and Deep Learning in Synthetic Biology: Key Architectures, Applications, and Challenges

Machine learning (ML), particularly deep learning (DL), has made rapid and substantial progress in synthetic biology in recent years. Biotechnological applications of biosystems, including pathways, enzymes, and whole cells, are being probed frequently with time. The intricacy and interconnectedness of biosystems make it challenging to design them with the desired properties. ML and DL have a synergy with synthetic biology. Synthetic biology can be employed to produce large data sets for training models (for instance, by utilizing DNA synthesis), and ML/DL models can be employed to inform design (for example, by generating new parts or advising unrivaled experiments to perform). This potential has recently been brought to light by research at the intersection of engineering biology and ML/DL through achievements like the design of novel biological components, best experimental design, automated analysis of microscopy data, protein structure prediction, and biomolecular implementations of ANNs (Artificial Neural Networks). I have divided this review into three sections. In the first section, I describe predictive potential and basics of ML along with myriad applications in synthetic biology, especially in engineering cells, activity of proteins, and metabolic pathways. In the second section, I describe fundamental DL architectures and their applications in synthetic biology. Finally, I describe different challenges causing hurdles in the progress of ML/DL and synthetic biology along with their solutions.

Predicting cell population-specific gene expression from genomic sequence

Most regulatory elements, especially enhancer sequences, are cell population-specific. One could even argue that a distinct set of regulatory elements is what defines a cell population. However, discovering which non-coding regions of the DNA are essential in which context, and as a result, which genes are expressed, is a difficult task. Some computational models tackle this problem by predicting gene expression directly from the genomic sequence. These models are currently limited to predicting bulk measurements and mainly make tissue-specific predictions. Here, we present a model that leverages single-cell RNA-sequencing data to predict gene expression. We show that cell population-specific models outperform tissue-specific models, especially when the expression profile of a cell population and the corresponding tissue are dissimilar. Further, we show that our model can prioritize GWAS variants and learn motifs of transcription factor binding sites. We envision that our model can be useful for delineating cell population-specific regulatory elements.

Canalizing cell fate by transcriptional repression

Precision in the establishment and maintenance of cellular identities is crucial for the development of multicellular organisms and requires tight regulation of gene expression. While extensive research has focused on understanding cell type-specific gene activation, the complex mechanisms underlying the transcriptional repression of alternative fates are not fully understood. Here, we provide an overview of the repressive mechanisms involved in cell fate regulation. We discuss the molecular machinery responsible for suppressing alternative fates and highlight the crucial role of sequence-specific transcription factors (TFs) in this process. Depletion of these TFs can result in unwanted gene expression and increased cellular plasticity. We suggest that these TFs recruit cell type-specific repressive complexes to their cis-regulatory elements, enabling them to modulate chromatin accessibility in a context-dependent manner. This modulation effectively suppresses master regulators of alternative fate programs and their downstream targets. The modularity and dynamic behavior of these repressive complexes enables a limited number of repressors to canalize and maintain major and minor cell fate decisions at different stages of development.

Cell-type-directed design of synthetic enhancers

Transcriptional enhancers act as docking stations for combinations of transcription factors and thereby regulate spatiotemporal activation of their target genes1. It has been a long-standing goal in the field to decode the regulatory logic of an enhancer and to understand the details of how spatiotemporal gene expression is encoded in an enhancer sequence. Here we show that deep learning models2-6, can be used to efficiently design synthetic, cell-type-specific enhancers, starting from random sequences, and that this optimization process allows detailed tracing of enhancer features at single-nucleotide resolution. We evaluate the function of fully synthetic enhancers to specifically target Kenyon cells or glial cells in the fruit fly brain using transgenic animals. We further exploit enhancer design to create 'dual-code' enhancers that target two cell types and minimal enhancers smaller than 50 base pairs that are fully functional. By examining the state space searches towards local optima, we characterize enhancer codes through the strength, combination and arrangement of transcription factor activator and transcription factor repressor motifs. Finally, we apply the same strategies to successfully design human enhancers, which adhere to enhancer rules similar to those of Drosophila enhancers. Enhancer design guided by deep learning leads to better understanding of how enhancers work and shows that their code can be exploited to manipulate cell states.

Mapping Cell Atlases at the Single-Cell Level

Recent advancements in single-cell technologies have led to rapid developments in the construction of cell atlases. These atlases have the potential to provide detailed information about every cell type in different organisms, enabling the characterization of cellular diversity at the single-cell level. Global efforts in developing comprehensive cell atlases have profound implications for both basic research and clinical applications. This review provides a broad overview of the cellular diversity and dynamics across various biological systems. In addition, the incorporation of machine learning techniques into cell atlas analyses opens up exciting prospects for the field of integrative biology.

Targeted design of synthetic enhancers for selected tissues in the Drosophila embryo

Enhancers control gene expression and have crucial roles in development and homeostasis1-3. However, the targeted de novo design of enhancers with tissue-specific activities has remained challenging. Here we combine deep learning and transfer learning to design tissue-specific enhancers for five tissues in the Drosophila melanogaster embryo: the central nervous system, epidermis, gut, muscle and brain. We first train convolutional neural networks using genome-wide single-cell assay for transposase-accessible chromatin with sequencing (ATAC-seq) datasets and then fine-tune the convolutional neural networks with smaller-scale data from in vivo enhancer activity assays, yielding models with 13% to 76% positive predictive value according to cross-validation. We designed and experimentally assessed 40 synthetic enhancers (8 per tissue) in vivo, of which 31 (78%) were active and 27 (68%) functioned in the target tissue (100% for central nervous system and muscle). The strategy of combining genome-wide and small-scale functional datasets by transfer learning is generally applicable and should enable the design of tissue-, cell type- and cell state-specific enhancers in any system.

Single-nucleus multi-omic profiling of human placental syncytiotrophoblasts identifies cellular trajectories during pregnancy

The human placenta has a vital role in ensuring a successful pregnancy. Despite the growing body of knowledge about its cellular compositions and functions, there has been limited research on the heterogeneity of the billions of nuclei within the syncytiotrophoblast (STB), a multinucleated entity primarily responsible for placental function. Here we conducted integrated single-nucleus RNA sequencing and single-nucleus ATAC sequencing analyses of human placentas from early and late pregnancy. Our findings demonstrate the dynamic heterogeneity and developmental trajectories of STB nuclei and their correspondence with human trophoblast stem cell (hTSC)-derived STB. Furthermore, we identified transcription factors associated with diverse STB nuclear lineages through their gene regulatory networks and experimentally confirmed their function in hTSC and trophoblast organoid-derived STBs. Together, our data provide insights into the heterogeneity of human STB and represent a valuable resource for interpreting associated pregnancy complications.

Inference of differentiation trajectories by transfer learning across biological processes

Stem cells differentiate into distinct fates by transitioning through a series of transcriptional states. Current computational approaches allow reconstruction of differentiation trajectories from single-cell transcriptomics data, but it remains unknown to what degree differentiation can be predicted across biological processes. Here, we use transfer learning to infer differentiation processes and quantify predictability in early embryonic development and adult hematopoiesis. Overall, we find that non-linear methods outperform linear approaches, and we achieved the best predictions with a custom variational autoencoder that explicitly models changes in transcriptional variance. We observed a high accuracy of predictions in embryonic development, but we found somewhat lower agreement with the real data in adult hematopoiesis. We demonstrate that this discrepancy can be explained by a higher degree of concordant transcriptional processes along embryonic differentiation compared with adult homeostasis. In summary, we establish a framework for quantifying and exploiting predictability of cellular differentiation trajectories.

MarsGT: Multi-omics analysis for rare population inference using single-cell graph transformer

Rare cell populations are key in neoplastic progression and therapeutic response, offering potential intervention targets. However, their computational identification and analysis often lag behind major cell types. To fill this gap, we introduce MarsGT: Multi-omics Analysis for Rare population inference using a Single-cell Graph Transformer. It identifies rare cell populations using a probability-based heterogeneous graph transformer on single-cell multi-omics data. MarsGT outperforms existing tools in identifying rare cells across 550 simulated and four real human datasets. In mouse retina data, it reveals unique subpopulations of rare bipolar cells and a Müller glia cell subpopulation. In human lymph node data, MarsGT detects an intermediate B cell population potentially acting as lymphoma precursors. In human melanoma data, it identifies a rare MAIT-like population impacted by a high IFN-I response and reveals the mechanism of immunotherapy. Hence, MarsGT offers biological insights and suggests potential strategies for early detection and therapeutic intervention of disease.

MICA: a multi-omics method to predict gene regulatory networks in early human embryos

Recent advances in single-cell omics have transformed characterisation of cell types in challenging-to-study biological contexts. In contexts with limited single-cell samples, such as the early human embryo inference of transcription factor-gene regulatory network (GRN) interactions is especially difficult. Here, we assessed application of different linear or non-linear GRN predictions to single-cell simulated and human embryo transcriptome datasets. We also compared how expression normalisation impacts on GRN predictions, finding that transcripts per million reads outperformed alternative methods. GRN inferences were more reproducible using a non-linear method based on mutual information (MI) applied to single-cell transcriptome datasets refined with chromatin accessibility (CA) (called MICA), compared with alternative network prediction methods tested. MICA captures complex non-monotonic dependencies and feedback loops. Using MICA, we generated the first GRN inferences in early human development. MICA predicted co-localisation of the AP-1 transcription factor subunit proto-oncogene JUND and the TFAP2C transcription factor AP-2γ in early human embryos. Overall, our comparative analysis of GRN prediction methods defines a pipeline that can be applied to single-cell multi-omics datasets in especially challenging contexts to infer interactions between transcription factor expression and target gene regulation.

Hold out the genome: a roadmap to solving the cis-regulatory code

Gene expression is regulated by transcription factors that work together to read cis-regulatory DNA sequences. The 'cis-regulatory code' - how cells interpret DNA sequences to determine when, where and how much genes should be expressed - has proven to be exceedingly complex. Recently, advances in the scale and resolution of functional genomics assays and machine learning have enabled substantial progress towards deciphering this code. However, the cis-regulatory code will probably never be solved if models are trained only on genomic sequences; regions of homology can easily lead to overestimation of predictive performance, and our genome is too short and has insufficient sequence diversity to learn all relevant parameters. Fortunately, randomly synthesized DNA sequences enable testing a far larger sequence space than exists in our genomes, and designed DNA sequences enable targeted queries to maximally improve the models. As the same biochemical principles are used to interpret DNA regardless of its source, models trained on these synthetic data can predict genomic activity, often better than genome-trained models. Here we provide an outlook on the field, and propose a roadmap towards solving the cis-regulatory code by a combination of machine learning and massively parallel assays using synthetic DNA.

Single-cell spatial multi-omics and deep learning dissect enhancer-driven gene regulatory networks in liver zonation

In the mammalian liver, hepatocytes exhibit diverse metabolic and functional profiles based on their location within the liver lobule. However, it is unclear whether this spatial variation, called zonation, is governed by a well-defined gene regulatory code. Here, using a combination of single-cell multiomics, spatial omics, massively parallel reporter assays and deep learning, we mapped enhancer-gene regulatory networks across mouse liver cell types. We found that zonation affects gene expression and chromatin accessibility in hepatocytes, among other cell types. These states are driven by the repressors TCF7L1 and TBX3, alongside other core hepatocyte transcription factors, such as HNF4A, CEBPA, FOXA1 and ONECUT1. To examine the architecture of the enhancers driving these cell states, we trained a hierarchical deep learning model called DeepLiver. Our study provides a multimodal understanding of the regulatory code underlying hepatocyte identity and their zonation state that can be used to engineer enhancers with specific activity levels and zonation patterns.

Modeling fragment counts improves single-cell ATAC-seq analysis

Single-cell ATAC sequencing coverage in regulatory regions is typically binarized as an indicator of open chromatin. Here we show that binarization is an unnecessary step that neither improves goodness of fit, clustering, cell type identification nor batch integration. Fragment counts, but not read counts, should instead be modeled, which preserves quantitative regulatory information. These results have immediate implications for single-cell ATAC sequencing analysis.

Direct androgen receptor control of sexually dimorphic gene expression in the mammalian kidney

Mammalian organs exhibit distinct physiology, disease susceptibility, and injury responses between the sexes. In the mouse kidney, sexually dimorphic gene activity maps predominantly to proximal tubule (PT) segments. Bulk RNA sequencing (RNA-seq) data demonstrated that sex differences were established from 4 and 8 weeks after birth under gonadal control. Hormone injection studies and genetic removal of androgen and estrogen receptors demonstrated androgen receptor (AR)-mediated regulation of gene activity in PT cells as the regulatory mechanism. Interestingly, caloric restriction feminizes the male kidney. Single-nuclear multiomic analysis identified putative cis-regulatory regions and cooperating factors mediating PT responses to AR activity in the mouse kidney. In the human kidney, a limited set of genes showed conserved sex-linked regulation, whereas analysis of the mouse liver underscored organ-specific differences in the regulation of sexually dimorphic gene expression. These findings raise interesting questions on the evolution, physiological significance, disease, and metabolic linkage of sexually dimorphic gene activity.

Predictive analyses of regulatory sequences with EUGENe

Deep learning has become a popular tool to study cis-regulatory function. Yet efforts to design software for deep-learning analyses in regulatory genomics that are findable, accessible, interoperable and reusable (FAIR) have fallen short of fully meeting these criteria. Here we present elucidating the utility of genomic elements with neural nets (EUGENe), a FAIR toolkit for the analysis of genomic sequences with deep learning. EUGENe consists of a set of modules and subpackages for executing the key functionality of a genomics deep learning workflow: (1) extracting, transforming and loading sequence data from many common file formats; (2) instantiating, initializing and training diverse model architectures; and (3) evaluating and interpreting model behavior. We designed EUGENe as a simple, flexible and extensible interface for streamlining and customizing end-to-end deep-learning sequence analyses, and illustrate these principles through application of the toolkit to three predictive modeling tasks. We hope that EUGENe represents a springboard towards a collaborative ecosystem for deep-learning applications in genomics research.

Gene regulatory network inference in the era of single-cell multi-omics

The interplay between chromatin, transcription factors and genes generates complex regulatory circuits that can be represented as gene regulatory networks (GRNs). The study of GRNs is useful to understand how cellular identity is established, maintained and disrupted in disease. GRNs can be inferred from experimental data - historically, bulk omics data - and/or from the literature. The advent of single-cell multi-omics technologies has led to the development of novel computational methods that leverage genomic, transcriptomic and chromatin accessibility information to infer GRNs at an unprecedented resolution. Here, we review the key principles of inferring GRNs that encompass transcription factor-gene interactions from transcriptomics and chromatin accessibility data. We focus on the comparison and classification of methods that use single-cell multimodal data. We highlight challenges in GRN inference, in particular with respect to benchmarking, and potential further developments using additional data modalities.

Butterfly eyespots exhibit unique patterns of open chromatin

Background: How the precise spatial regulation of genes is correlated with spatial variation in chromatin accessibilities is not yet clear. Previous studies that analysed chromatin from homogenates of whole-body parts of insects found little variation in chromatin accessibility across those parts, but single-cell studies of Drosophila brains showed extensive spatial variation in chromatin accessibility across that organ. In this work we studied the chromatin accessibility of butterfly wing tissue fated to differentiate distinct colors and patterns in pupal wings of Bicyclus anynana. Methods: We dissected small eyespot and adjacent control tissues from 3h pupae and performed ATAC-Seq to identify the chromatin accessibility differences between different sections of the wings. Results: We observed that three dissected wing regions showed unique chromatin accessibilities. Open chromatin regions specific to eyespot color patterns were highly enriched for binding motifs recognized by Suppressor of Hairless (Su(H)), Krüppel (Kr), Buttonhead (Btd) and Nubbin (Nub) transcription factors. Genes in the vicinity of the eyespot-specific open chromatin regions included those involved in wound healing and SMAD signal transduction pathways, previously proposed to be involved in eyespot development. Conclusions: We conclude that eyespot and non-eyespot tissue samples taken from the same wing have distinct patterns of chromatin accessibility, possibly driven by the eyespot-restricted expression of potential pioneer factors, such as Kr.

Transcription factor stoichiometry, motif affinity and syntax regulate single-cell chromatin dynamics during fibroblast reprogramming to pluripotency

Ectopic expression of OCT4, SOX2, KLF4 and MYC (OSKM) transforms differentiated cells into induced pluripotent stem cells. To refine our mechanistic understanding of reprogramming, especially during the early stages, we profiled chromatin accessibility and gene expression at single-cell resolution across a densely sampled time course of human fibroblast reprogramming. Using neural networks that map DNA sequence to ATAC-seq profiles at base-resolution, we annotated cell-state-specific predictive transcription factor (TF) motif syntax in regulatory elements, inferred affinity- and concentration-dependent dynamics of Tn5-bias corrected TF footprints, linked peaks to putative target genes, and elucidated rewiring of TF-to-gene cis-regulatory networks. Our models reveal that early in reprogramming, OSK, at supraphysiological concentrations, rapidly open transient regulatory elements by occupying non-canonical low-affinity binding sites. As OSK concentration falls, the accessibility of these transient elements decays as a function of motif affinity. We find that these OSK-dependent transient elements sequester the somatic TF AP-1. This redistribution is strongly associated with the silencing of fibroblast-specific genes within individual nuclei. Together, our integrated single-cell resource and models reveal insights into the cis-regulatory code of reprogramming at unprecedented resolution, connect TF stoichiometry and motif syntax to diversification of cell fate trajectories, and provide new perspectives on the dynamics and role of transient regulatory elements in somatic silencing.

SCALA: A complete solution for multimodal analysis of single-cell Next Generation Sequencing data

Analysis and interpretation of high-throughput transcriptional and chromatin accessibility data at single-cell (sc) resolution are still open challenges in the biomedical field. The existence of countless bioinformatics tools, for the different analytical steps, increases the complexity of data interpretation and the difficulty to derive biological insights. In this article, we present SCALA, a bioinformatics tool for analysis and visualization of single-cell RNA sequencing (scRNA-seq) and Assay for Transposase-Accessible Chromatin using sequencing (scATAC-seq) datasets, enabling either independent or integrative analysis of the two modalities. SCALA combines standard types of analysis by integrating multiple software packages varying from quality control to the identification of distinct cell populations and cell states. Additional analysis options enable functional enrichment, cellular trajectory inference, ligand-receptor analysis, and regulatory network reconstruction. SCALA is fully parameterizable, presenting data in tabular format and producing publication-ready visualizations. The different available analysis modules can aid biomedical researchers in exploring, analyzing, and visualizing their data without any prior experience in coding. We demonstrate the functionality of SCALA through two use-cases related to TNF-driven arthritic mice, handling both scRNA-seq and scATAC-seq datasets. SCALA is developed in R, Shiny and JavaScript and is mainly available as a standalone version, while an online service of more limited capacity can be found at http://scala.pavlopouloslab.info or https://scala.fleming.gr.

The evolution and mutational robustness of chromatin accessibility in Drosophila

BACKGROUND

The evolution of genomic regulatory regions plays a critical role in shaping the diversity of life. While this process is primarily sequence-dependent, the enormous complexity of biological systems complicates the understanding of the factors underlying regulation and its evolution. Here, we apply deep neural networks as a tool to investigate the sequence determinants underlying chromatin accessibility in different species and tissues of Drosophila.

RESULTS

We train hybrid convolution-attention neural networks to accurately predict ATAC-seq peaks using only local DNA sequences as input. We show that our models generalize well across substantially evolutionarily diverged species of insects, implying that the sequence determinants of accessibility are highly conserved. Using our model to examine species-specific gains in accessibility, we find evidence suggesting that these regions may be ancestrally poised for evolution. Using in silico mutagenesis, we show that accessibility can be accurately predicted from short subsequences in each example. However, in silico knock-out of these sequences does not qualitatively impair classification, implying that accessibility is mutationally robust. Subsequently, we show that accessibility is predicted to be robust to large-scale random mutation even in the absence of selection. Conversely, simulations under strong selection demonstrate that accessibility can be extremely malleable despite its robustness. Finally, we identify motifs predictive of accessibility, recovering both novel and previously known motifs.

CONCLUSIONS

These results demonstrate the conservation of the sequence determinants of accessibility and the general robustness of chromatin accessibility, as well as the power of deep neural networks to explore fundamental questions in regulatory genomics and evolution.

Deep flanking sequence engineering for efficient promoter design using DeepSEED

Designing promoters with desirable properties is essential in synthetic biology. Human experts are skilled at identifying strong explicit patterns in small samples, while deep learning models excel at detecting implicit weak patterns in large datasets. Biologists have described the sequence patterns of promoters via transcription factor binding sites (TFBSs). However, the flanking sequences of cis-regulatory elements, have long been overlooked and often arbitrarily decided in promoter design. To address this limitation, we introduce DeepSEED, an AI-aided framework that efficiently designs synthetic promoters by combining expert knowledge with deep learning techniques. DeepSEED has demonstrated success in improving the properties of Escherichia coli constitutive, IPTG-inducible, and mammalian cell doxycycline (Dox)-inducible promoters. Furthermore, our results show that DeepSEED captures the implicit features in flanking sequences, such as k-mer frequencies and DNA shape features, which are crucial for determining promoter properties.

Principles and challenges of modeling temporal and spatial omics data

Studies with temporal or spatial resolution are crucial to understand the molecular dynamics and spatial dependencies underlying a biological process or system. With advances in high-throughput omic technologies, time- and space-resolved molecular measurements at scale are increasingly accessible, providing new opportunities to study the role of timing or structure in a wide range of biological questions. At the same time, analyses of the data being generated in the context of spatiotemporal studies entail new challenges that need to be considered, including the need to account for temporal and spatial dependencies and compare them across different scales, biological samples or conditions. In this Review, we provide an overview of common principles and challenges in the analysis of temporal and spatial omics data. We discuss statistical concepts to model temporal and spatial dependencies and highlight opportunities for adapting existing analysis methods to data with temporal and spatial dimensions.

Hormonal gatekeeping via the blood-brain barrier governs caste-specific behavior in ants

Here, we reveal an unanticipated role of the blood-brain barrier (BBB) in regulating complex social behavior in ants. Using scRNA-seq, we find localization in the BBB of a key hormone-degrading enzyme called juvenile hormone esterase (Jhe), and we show that this localization governs the level of juvenile hormone (JH3) entering the brain. Manipulation of the Jhe level reprograms the brain transcriptome between ant castes. Although ant Jhe is retained and functions intracellularly within the BBB, we show that Drosophila Jhe is naturally extracellular. Heterologous expression of ant Jhe into the Drosophila BBB alters behavior in fly to mimic what is seen in ants. Most strikingly, manipulation of Jhe levels in ants reprograms complex behavior between worker castes. Our study thus uncovers a remarkable, potentially conserved role of the BBB serving as a molecular gatekeeper for a neurohormonal pathway that regulates social behavior.

Lola-I is a promoter pioneer factor that establishes de novo Pol II pausing during development

While the accessibility of enhancers is dynamically regulated during development, promoters tend to be constitutively accessible and poised for activation by paused Pol II. By studying Lola-I, a Drosophila zinc finger transcription factor, we show here that the promoter state can also be subject to developmental regulation independently of gene activation. Lola-I is ubiquitously expressed at the end of embryogenesis and causes its target promoters to become accessible and acquire paused Pol II throughout the embryo. This promoter transition is required but not sufficient for tissue-specific target gene activation. Lola-I mediates this function by depleting promoter nucleosomes, similar to the action of pioneer factors at enhancers. These results uncover a level of regulation for promoters that is normally found at enhancers and reveal a mechanism for the de novo establishment of paused Pol II at promoters.

SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks

Joint profiling of chromatin accessibility and gene expression in individual cells provides an opportunity to decipher enhancer-driven gene regulatory networks (GRNs). Here we present a method for the inference of enhancer-driven GRNs, called SCENIC+. SCENIC+ predicts genomic enhancers along with candidate upstream transcription factors (TFs) and links these enhancers to candidate target genes. To improve both recall and precision of TF identification, we curated and clustered a motif collection with more than 30,000 motifs. We benchmarked SCENIC+ on diverse datasets from different species, including human peripheral blood mononuclear cells, ENCODE cell lines, melanoma cell states and Drosophila retinal development. Next, we exploit SCENIC+ predictions to study conserved TFs, enhancers and GRNs between human and mouse cell types in the cerebral cortex. Finally, we use SCENIC+ to study the dynamics of gene regulation along differentiation trajectories and the effect of TF perturbations on cell state. SCENIC+ is available at scenicplus.readthedocs.io .