Biology subjects

Zeitlinger, J.

Publications and source records attributed to Zeitlinger, J..

16 recordsLinked to original sources

bioRxiv2026

Temporal regulation of a spatial patterning factor in Drosophila neurogenesis

A central question in neurobiology is how the transient programs that pattern neural progenitors are translated into the enormous, stable diversity of neuronal types. Spatial and temporal cues act only briefly, yet each neuron's identity is defined and maintained for life by terminal selector transcription factors (TFs). How a neuron's developmental origin is read out into a particular selector code remains poorly understood. Some current models propose that spatial and temporal origins are inherited independently through separate selectors. We show instead that, in the Drosophila optic lobe, the same selector can be activated by different patterning axes through physically distinct enhancers, even within the same lineage. Visual system homeobox (Vsx1) spatially patterns a central neuroepithelial domain and later acts as a terminal selector in dozens of neuronal types, most originating exclusively from that domain. However, in Dm2 neurons that are produced from every domain, it is regulated not by neuroepithelial Vsx1 but by the neuroblast temporal TF BarH1, through an enhancer distinct from its domain-specific ones. Combining in vivo reporters with sequence-to-accessibility deep-learning models, we identify and disrupt the key binding sites in this enhancer, impairing its Dm2-specific activity. Reciprocally, the temporal TF Homeobrain (Hbn) acts as a terminal selector in the related neuron Mi21 independently of its neuroblast temporal window: its expression in these late-born neurons is instead placed under dorsoventral spatial control. Patterning inputs therefore need not be partitioned across separate selectors but converge combinatorially on the modular enhancers of shared ones, revealing a cis-regulatory logic that re-encodes this limited set of inputs into vast neuronal diversity.

neuroscience↗

bioRxiv2025

High-resolution binding data of TFIID and cofactors show promoter-specific differences in vivo

TFIID is instrumental in recognizing promoter sequences and initiating transcription, yet a cohesive understanding of how this complex interacts with and functions at different promoter types in vivo is still lacking. Here, we employed ChIP-nexus to capture high-resolution binding footprints of all Drosophila TFIID subunits across the genome. These footprints reveal TFIID sub-modules whose DNA contacts suggest new structural details. At different promoter types, the footprints of the TAFs are very similar, suggesting the presence of engaged TFIID across all promoters. In contrast, the binding profile of TBP is promoter-specific, enabling us to identify TATA, DPR, and TCT/housekeeping promoters de novo, along with their underlying core promoter elements. Notably, our data point to NC2 being specific for TBP binding at the TATA box and suggest that TATA promoters show both TAF-dependent and TAF-independent initiation in vivo. These data suggest a model for the increased burst size observed at TATA promoters and provide a comprehensive resource for linking structural and biochemical results to in vivo data.

genomics↗

bioRxiv2025

Widespread low-affinity motifs enhance chromatin accessibility and regulatory potential in mESCs

Low-affinity transcription factor (TF) motifs are an important element of the cis-regulatory code, yet they are notoriously difficult to map and mechanistically incompletely understood, limiting our ability to interpret non-coding variation in development, evolution, and disease. Here we investigate their role in pioneering and leverage sequence-to-profile models of chromatin accessibility in mouse embryonic stem cells to reliably map and interpret low-affinity motifs across the genome. We find that low-affinity motifs have outsized effects by cooperating with nearby motifs through intra-nucleosomal soft syntax. By modeling nucleosome-mediated cooperativity with a kinetic model, we discover and validate that pioneer cooperativity makes a motif operate at higher pioneering ranges across changing TF concentrations, thereby raising the regulatory potential. These results show that low-affinity motifs can be accurately mapped, shape the properties of developmental enhancers and likely play a widespread role in fine-tuning enhancers during evolution.

genomics↗

bioRxiv2025

GAME: Genomic API for Model Evaluation

The rapid expansion of genomics datasets and the application of machine learning has produced sequence-to-activity genomics models with ever-expanding capabilities. However, benchmarking these models on practical applications has been challenging because individual projects evaluate their models in ad hoc ways, and there is substantial heterogeneity of both model architectures and benchmarking tasks. To address this challenge, we have created GAME, a system for large-scale, community-led standardized model benchmarking on user-defined evaluation tasks. We borrow concepts from the Application Programming Interface (API) paradigm to allow for seamless communication between pre-trained models and benchmarking tasks, ensuring consistent evaluation protocols. Because all models and benchmarks are inherently compatible in this framework, the continual addition of new models and new benchmarks is easy. We also developed a Matcher module powered by a large language model (LLM) to automate ambiguous task alignment between benchmarks and models. Containerization of these modules enhances reproducibility and facilitates the deployment of models and benchmarks across computing platforms. By focusing on predicting underlying biochemical phenomena (e.g. gene expression, open chromatin, DNA binding), we ensure that tasks remain technology-independent. We provide examples of benchmarks and models implementing this framework, and anticipate that the community will contribute their own, leading to an ever-expanding and evolving set of models and evaluation tasks. This resource will accelerate genomics research by illuminating the best models for a given task, motivating novel functional genomic benchmarks, and providing a more nuanced understanding of model abilities.

bioinformatics↗

bioRxiv2025

PISA: a versatile interpretation tool for visualizing cis-regulatory rules in genomic data

Sequence-to-function neural networks learn cis-regulatory sequence rules driving many types of genomic data. Interpreting these models to relate the sequence rules to underlying biological processes remains challenging, especially for complex genomic readouts such as MNase-seq, which maps nucleosome occupancy but is confounded by experimental bias. We introduce pairwise influence by sequence attribution (PISA), an interpretation tool that combinatorially decodes which bases contributed to the readout at a specific genomic coordinate. PISA visualizes the effects of transcription factor motifs, detects undiscovered motifs with complex contribution patterns, and reveals experimental biases. By learning the bias for MNase-seq, PISA enables unprecedented nucleosome prediction models, allowing the de novo discovery of nucleosome-positioning motifs and their longrange chromatin effects, as well as the design of sequences with altered nucleosome configurations. These results show that PISA is a versatile tool that expands our ability to train and interpret sequence-to-function neural networks on genomics data and understand the underlying cis-regulatory code.

genomics↗

bioRxiv2024

Gene regulatory network co-option is sufficient to induce a morphological novelty in Drosophila

Identifying the molecular origins by which new morphological structures evolve is one of the long standing problems in evolutionary biology. To date, vanishingly few examples provide a compelling account of how new morphologies were initially formed, thereby limiting our understanding of how diverse forms of life derived their complex features. Here, we provide evidence that the large projections on the Drosophila eugracilis phallus that are implicated in sexual conflict have evolved through co-option of the trichome genetic network. These unicellular apical projections on the phallus postgonal sheath are reminiscent of trichomes that cover the Drosophila body but are up to 20-fold larger in size. During their development, they express the transcription factor Shavenbaby, the master regulator of the trichome network. Consistent with the co-option of the Shavenbaby network during the evolution of the D. eugracilis projections, somatic mosaic CRISPR/Cas9 mutagenesis shows that shavenbaby is necessary for their proper length. Moreover, mis-expression of Shavenbaby in the sheath of D. melanogaster, a naive species that lacks these extensions, is sufficient to induce small trichomes. These induced extensions rely on a genetic network that is shared to a large extent with the D. eugracilis projections, indicating its co-option but also some genetic rewiring. Thus, by leveraging a genetically tractable evolutionarily novelty, our work shows that the trichome-forming network is flexible enough that it can be co-opted in a new context, and subsequently refined to produce unique apical projections that are barely recognizable compared to their simpler ancestral beginnings.

evolutionary biology↗

bioRxiv2024

Interpretable deep learning reveals the sequence rules of Hippo signaling

The response to signaling pathways is highly context-specific, and identifying the transcription factors and mechanisms that are responsible is very challenging. Using the Hippo pathway in mouse trophoblast stem cells as a model, we show here that this information is encoded in cis-regulatory sequences and can be learned from high-resolution binding data of signaling transcription factors. Using interpretable deep learning, we show that the binding levels of TEAD4 and YAP1 are enhanced in a distance-dependent manner by cell type-specific transcription factors, including TFAP2C. We also discovered that strictly spaced Tead double motifs are widespread highly active canonical response elements that mediate cooperativity by promoting labile TEAD4 protein-protein interactions on DNA. These syntax rules and mechanisms apply genome-wide and allow us to predict how small sequence changes alter the activity of enhancers in vivo. This illustrates the power of interpretable deep learning to decode canonical and cell type-specific sequence rules of signaling pathways. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=183 SRC="FIGDIR/small/580842v1_ufig1.gif" ALT="Figure 1"> View larger version (48K): org.highwire.dtl.DTLVardef@1ee22e9org.highwire.dtl.DTLVardef@1362c82org.highwire.dtl.DTLVardef@1a42512org.highwire.dtl.DTLVardef@17dd619_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

bioRxiv2023

De novo distillation of thermodynamic affinity from deep learning regulatory sequence models of in vivo protein-DNA binding

Transcription factors (TF) are proteins that bind DNA in a sequence-specific manner to regulate gene transcription. Despite their unique intrinsic sequence preferences, in vivo genomic occupancy profiles of TFs differ across cellular contexts. Hence, deciphering the sequence determinants of TF binding, both intrinsic and context-specific, is essential to understand gene regulation and the impact of regulatory, non-coding genetic variation. Biophysical models trained on in vitro TF binding assays can estimate intrinsic affinity landscapes and predict occupancy based on TF concentration and affinity. However, these models cannot adequately explain context-specific, in vivo binding profiles. Conversely, deep learning models, trained on in vivo TF binding assays, effectively predict and explain genomic occupancy profiles as a function of complex regulatory sequence syntax, albeit without a clear biophysical interpretation. To reconcile these complementary models of in vitro and in vivo TF binding, we developed Affinity Distillation (AD), a method that extracts thermodynamic affinities de-novo from deep learning models of TF chromatin immunoprecipitation (ChIP) experiments by marginalizing away the influence of genomic sequence context. Applied to neural networks modeling diverse classes of yeast and mammalian TFs, AD predicts energetic impacts of sequence variation within and surrounding motifs on TF binding as measured by diverse in vitro assays with superior dynamic range and accuracy compared to motif-based methods. Furthermore, AD can accurately discern affinities of TF paralogs. Our results highlight thermodynamic affinity as a key determinant of in vivo binding, suggest that deep learning models of in vivo binding implicitly learn high-resolution affinity landscapes, and show that these affinities can be successfully distilled using AD. This new biophysical interpretation of deep learning models enables high-throughput in silico experiments to explore the influence of sequence context and variation on both intrinsic affinity and in vivo occupancy.

genomics↗

bioRxiv2022

Chromatin accessibility is a two-tier process regulated by transcription factor pioneering and enhancer activation

Chromatin accessibility is integral to the process by which transcription factors (TFs) read out cis-regulatory DNA sequences, but it is difficult to differentiate between TFs that drive accessibility and those that do not. Deep learning models that learn complex sequence rules provide an unprecedented opportunity to dissect this problem. Using zygotic genome activation in the Drosophila embryo as a model, we generated high-resolution TF binding and chromatin accessibility data, analyzed the data with interpretable deep learning, and performed genetic experiments for validation. We uncover a clear hierarchical relationship between the pioneer TF Zelda and the TFs involved in axis patterning. Zelda consistently pioneers chromatin accessibility proportional to motif affinity, while patterning TFs augment chromatin accessibility in sequence contexts in which they mediate enhancer activation. We conclude that chromatin accessibility occurs in two phases: one through pioneering, which makes enhancers accessible but not necessarily active, and a second when the correct combination of transcription factors leads to enhancer activation.

genomics↗

bioRxiv2022

Short tandem repeats bind transcription factors to tune eukaryotic gene expression

Short tandem repeats (STRs) are enriched in eukaryotic cis-regulatory elements and their polymorphisms alter gene expression, yet how they regulate transcription remains unknown. We find that STRs can modulate transcription factor (TF)-DNA affinities and on rates by up to 70-fold by directly binding TF DNA-binding domains, with energetic impacts approaching or exceeding mutations to consensus sites. STRs maximize the number of weakly preferred microstates near target sites, thereby increasing TF density near motifs to speed target search. Confirming that STRs also impact TF binding in cells, neural networks trained only on in vivo occupancies predict identical effects to those observed in vitro. Approximately 90% of TFs preferentially bind STRs that need not resemble known motifs, providing a novel cis-regulatory mechanism to target TFs to cognate sites.

biophysics↗

bioRxiv2022

Lola-I is a developmentally regulated promoter pioneer factor

While enhancers are often regulated at the level of accessibility by pioneer factors, promoters tend to be constitutively accessible and poised for activation by paused Pol II -- thus are often not considered as sites of developmental regulation. Here we show that the accessibility of promoters and the acquisition of paused Pol II can be subject to developmental regulation by pioneer factors. We show that Lola-I, a Drosophila zinc finger transcription factor, 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 expression. Lola-I mediates this function by binding to the edges of the promoter nucleosomes, which leads to their depletion, similar to the action of pioneer factors at enhancers. These results uncover a level of regulation for promoters that is normally found at enhancers, providing further evidence that promoters and enhancers display unexpectedly similar characteristics.

genomics↗

bioRxiv2021

The SAGA core module is critical during Drosophila oogenesis and is broadly recruited to promoters

The Spt/Ada-Gcn5 Acetyltransferase (SAGA) coactivator complex has multiple modules with different enzymatic and non-enzymatic functions. How each module contributes to gene activation in specific biological contexts is not well understood. Here we analyzed the role of the non-enzymatic core module during Drosophila oogenesis. We show that depletion of several SAGA-specific subunits belonging to the core module blocked egg chamber development during mid-oogenesis stages, resulting in stronger phenotypes than those obtained after depletion of SAGAs histone acetyltransferase module or deubiquitination module. These results, as well as additional genetic analyses pointing to an interaction with TBP, suggested a differential role of SAGA modules at different promoter types. However, SAGA subunits co-occupied all promoter types of active genes in ChIP-seq and ChIP-nexus experiments. Thus, the SAGA complex appears to occupy promoters in its entirety, consistent with the strong biochemical integrity of the complex. The high-resolution genomic binding profiles are congruent with SAGA recruitment by activators upstream of the start site, and retention on chromatin by interactions with modified histones downstream of the start site. The stronger genetic requirement of the core module during oogenesis may therefore be explained through its interaction with TBP or its role in recruiting the enzymatic modules to the promoter. We propose the handyman principle, which posits that a distinct genetic requirement for specific components may conceal the fact that the entire complex is physically present. Author SummaryEmbryonic development critically relies on the differential expression of genes in different tissues. This involves the dynamic interplay between DNA, sequence-specific transcription factors, coactivators and chromatin remodelers which guide the transcription machinery to the appropriate promoters for productive transcription. To understand how this happens at the molecular level, we need to understand when and how coactivator complexes such as SAGA function. SAGA consists of multiple modules with well characterized enzymatic functions. This study shows that the non-enzymatic core module of SAGA is required for Drosophila oogenesis, while the enzymatic functions are largely dispensable. Despite this differential requirement, SAGA subunits appear to be broadly recruited to all promoter types, consistent with the biochemical integrity of the complex. These results suggest that genetic requirements and physical organization do not always go hand in hand.

genomics↗

bioRxiv2020

TATA and paused promoters active in differentiated tissues have distinct expression characteristics

Core promoter types differ in the extent to which RNA polymerase II (Pol II) pauses after initiation, but how this difference affects their tissue-specific gene expression characteristics is not well understood. While promoters with Pol II pausing elements are active at all stages of development, TATA promoters are highly active in differentiated tissues. We therefore used a genomics approach on late-stage Drosophila embryos to analyze the properties of promoter types. Using tissue-specific Pol II ChIP-seq, we found that paused promoters have high levels of paused Pol II throughout the embryo, even in tissues where the gene is not expressed, while TATA promoters only show Pol II occupancy when the gene is active. This difference between promoter types is associated with different chromatin accessibility in ATAC-seq data and different expression characteristics in single-cell RNA data. The results suggest that promoter types have optimized different promoter properties: paused promoters show more consistent expression when active, while TATA promoters have lower background expression when inactive. We propose that tissue-specific effector genes have evolved to use two different strategies for their differential expression across tissues.

genomics↗

bioRxiv2019

Deep learning at base-resolution reveals motif syntax of the cis-regulatory code

The arrangement of transcription factor (TF) binding motifs (syntax) is an important part of the cis-regulatory code, yet remains elusive. We introduce a deep learning model, BPNet, that uses DNA sequence to predict base-resolution ChIP-nexus binding profiles of pluripotency TFs. We develop interpretation tools to learn predictive motif representations and identify soft syntax rules for cooperative TF binding interactions. Strikingly, Nanog preferentially binds with helical periodicity, and TFs often cooperate in a directional manner, which we validate using CRISPR-induced point mutations. Our model represents a powerful general approach to uncover the motifs and syntax of cis-regulatory sequences in genomics data. HighlightsO_LIThe neural network BPNet accurately predicts TF binding data at base-resolution. C_LIO_LIModel interpretation discovers TF motifs and TF interactions dependent on soft syntax. C_LIO_LIMotifs for Nanog and partners are preferentially spaced at [~]10.5 bp periodicity. C_LIO_LIDirectional cooperativity is validated: Sox2 enhances Nanog binding, but not vice versa. C_LI

genomics↗

bioRxiv2019

An atlas of transcription factors expressed in the Drosophila melanogaster pupal terminalia

During development, transcription factors and signaling molecules govern gene regulatory networks to direct the formation of unique morphologies. As changes in gene regulatory networks are often implicated in morphological evolution, mapping transcription factor landscapes is important, especially in tissues that undergo rapid evolutionary change. The terminalia (genital and anal structures) of Drosophila melanogaster and its close relatives exhibit dramatic changes in morphology between species. While previous studies have found network components important for patterning the larval genital disc, the networks governing adult structures during pupal development have remained uncharted. Here, we performed RNA-seq in whole Drosophila melanogaster terminalia followed by in situ hybridization for 100 highly expressed transcription factors during pupal development. We find that the terminalia is highly patterned during pupal stages and that specific transcription factors mark separate structures and substructures. Our results are housed online in a searchable database (flyterminalia.pitt.edu) where they can serve as a resource for the community. This work lays a foundation for future investigations into the gene regulatory networks governing the development and evolution of Drosophila terminalia.\n\nSummaryWe performed RNA-seq in whole Drosophila melanogaster terminalia (genitalia and analia) followed by in situ hybridization for 100 highly expressed transcription factors during pupal development. We find that the pupal terminalia is highly patterned with specific transcription factors marking separate structures and substructures. Our results are housed online in a searchable database (flyterminalia.pitt.edu) where they can serve as a resource for the community. This work lays a foundation for future investigations into the gene regulatory networks governing the development and evolution of Drosophila terminalia.

genetics↗

bioRxiv2018

High-quality genome assemblies of 15 Drosophila species generated using Nanopore sequencing

The Drosophila genus is a unique group containing a wide range of species that occupy diverse ecosystems. In addition to the most widely studied species, Drosophila melanogaster, many other members in this genus also possess a well-developed set of genetic tools. Indeed, high-quality genomes exist for several species within the genus, facilitating studies of the function and evolution of cis-regulatory regions and proteins by allowing comparisons across at least 50 million years of evolution. Yet, the available genomes still fail to capture much of the substantial genetic diversity within the Drosophila genus. We have therefore tested protocols to rapidly and inexpensively sequence and assemble the genome from any Drosophila species using single-molecule sequencing technology from Oxford Nanopore. Here, we use this technology to present high-quality genome assemblies of 15 Drosophila species: 10 of the 12 originally sequenced Drosophila species (ananassae, erecta, mojavensis, persimilis, pseudoobscura, sechellia, simulans, virilis, willistoni, and yakuba), four additional species that had previously reported assemblies (biarmipes, bipectinata, eugracilis, and mauritiana), and one novel assembly (triauraria). Genomes were generated from an average of 29x depth-of-coverage data that after assembly resulted in an average contig N50 of 4.4 Mb. Subsequent alignment of contigs from the published reference genomes demonstrates that our assemblies could be used to close over 60% of the gaps present in the currently published reference genomes. Importantly, the materials and reagents cost for each genome was approximately $1,000 (USD). This study demonstrates the power and cost-effectiveness of long-read sequencing for genome assembly in Drosophila and provides a framework for the affordable sequencing and assembly of additional Drosophila genomes.

genomics↗