Nature

Panarthropod vision exhibits extraordinary morphological and functional diversity, yet the sensory biology of tardigrades--microscopic extremophiles renowned for their resilience--remains poorly understood. In the model tardigrade Hypsibius exemplaris, we uncover an unprecedented expansion of opsin genes, with over 100 paralogs constituting the largest known opsin repertoire in any animal. Paradoxically, the visual system is structurally minimalist: a paired, inverse pigment-cup ocellus embedded within the brain lobes, forming a single-pixel, dual-receptor organ. Integrating genomic, phylogenetic, molecular expression, and ultrastructural analyses, we show that directional vision relies on a single rhabdomeric opsin (He-R-Opsin-V), localized to microvilli of the rhabdomeric cell. A ciliary photoreceptor with a lamellated cilium co-expresses two ciliary opsins (He-C-Opsin-1 and -2), suggesting non-visual light detection. These and other non-visual opsins are differentially expressed in the brain, gut, storage cells, and peripheral tissues, implicating them in circadian regulation, neuromodulation, ecdysis, digestion, and environmental sensing. Crucially, the eye is an internalized epidermal vesicle, not a cerebral derivative, challenging long-standing assumptions about its evolutionary origin. These findings reveal how extreme miniaturization drives sensory system simplification in visual organs while enabling parallel evolutionary innovation in non-visual photoreception. This study establishes a new paradigm for sensory evolution in microscale animals.

Pathogens hijack macrophages by triggering pathological cAMP surges that block phagolysosomal killing--a defect mirrored in phagocytes from refractory colitis. We identify a host-encoded, pathogen-specific surge-protector comprised of a three-protein toggle: The innate sensor NOD2 binds and masks an evolutionarily conserved motif in GIV that activates trimeric-GTPase Gi, enforcing a biphasic surge-to-plunge cAMP-program: early, NOD2*GIV assembly permits a brief, tolerogenic cAMP rise, whereas subsequent GIV*Gi engagement collapses cAMP to drive phagolysosomal fusion and microbial clearance. Structural, biochemical, and ultrastructural analyses reveal how molecular toggling imposes precise spatial and temporal control. Pharmacogenomic perturbations pinpoint cAMP-PKA hyperactivation as the defining lesion in GIV-deficient macrophages. Functional studies in primary macrophages and human gut organoid co-cultures show that toggling the NOD2*GIV*Gi-axis is necessary and sufficient to convert tolerant macrophages into microbicidal machines that preserve mucosal barrier integrity. These findings uncover a druggable cAMP-control pathway with therapeutic promise in colitis. GRAPHIC ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/715116v1_ufig1.gif" ALT="Figure 1"> View larger version (70K): org.highwire.dtl.DTLVardef@6f7c6corg.highwire.dtl.DTLVardef@151d439org.highwire.dtl.DTLVardef@1441b12org.highwire.dtl.DTLVardef@4d76e3_HPS_FORMAT_FIGEXP M_FIG C_FIG eTOC BlurbPathogens hijack macrophages by inducing cAMP surges that help them evade clearance. Anandachar et al. identify a host "toggle switch" in which NOD2 and G proteins compete for GIV, driving a rapid and robust surge-to-plunge transition in cAMP. This temporal switch limits tolerogenic signaling, restores microbial clearance and barrier integrity, and unveils a targetable host pathway in infection and IBD. HighlightsO_LIPathogens exploit cAMP surges in macrophages to block phagolysosomal killing of microbes C_LIO_LIGIV acts as a molecular "toggle" linking NOD2 sensing to Gi-mediated cAMP control C_LIO_LIStructural and mutagenesis studies reveal mutually exclusive binding of NOD2 and Gi to GIV C_LIO_LIPharmacogenomic perturbations pinpoint PKA, not EPAC, as the critical downstream effector C_LIO_LIOrganoid co-cultures show NOD2-GIV-PKA crosstalk safeguards microbial clearance and gut barrier integrity C_LI

Organisms must regulate metabolic resources such as oxygen (O2) and nutrients despite environmental variability and the energetic costs of their own actions1-3. Such regulation can occur reactively, through homeostatic corrections of recent imbalances, or predictively, through allostatic adjustments that anticipate future demand4,5. Predictive regulation is particularly important because metabolic resources often continue to be consumed for seconds to minutes after motor actions cease as tissues repay incurred costs, making it advantageous to prevent depletion before it occurs6. However, the cellular and circuit mechanisms for allostatic control remain largely unknown5,7,8. Using whole-brain neuronal and astroglial imaging and O2 measurements in behaving zebrafish, we identified a noradrenergic-astroglial circuit that detects, anticipates, and prevents internal O2 depletion. We found that swimming exacerbated internal hypoxia with a multi-second delay, but behavioral adaptations occurred before such self-generated hypoxia manifested, suggesting predictive control, confirmed using computational modeling. Noradrenergic neurons in the nucleus of the solitary tract directly detected brain hypoxia and received efference copies of swimming actions; these inputs summed at the level of membrane voltage to increase spiking and norepinephrine release when actions and resource scarcity co-occurred. Astroglia integrated noradrenergic input into prolonged Ca2+ elevation that tracked the O2 cost of recent actions and thereby predicted O2 debt relative to O2 availability, rising [~]8 s before O2 fell. This astroglial prediction reorganized brain-wide activity to suppress locomotion and promote respiration, preempting O2 depletion. Silencing noradrenergic neurons or astroglial signaling abolished these hypoxia coping behaviors, whereas selective activation evoked them. This neuronal-astroglial mechanism constitutes a predictive control system that integrates physiological state with behavioral intent to avert metabolic crisis, revealing a cellular substrate for proactive energy management.

How can humans remember and imagine without vision? Mental time travel, the ability to re-experience past events and envision future ones, is widely assumed to rely on visual imagery and the construction of mental scenes. Blindness provides a critical test of this assumption. Across behavioral interviews, language analyses, and multimodal neuroimaging in congenitally blind, late-blind, and sighted individuals, we show that blind individuals, even those blind from birth, mentally time travel as vividly as sighted people, but construct their inner worlds differently. Sighted participants relied on perceptual detail and activated classic scene-processing regions, whereas blind participants emphasized thoughts and emotions and recruited reorganized occipital cortex. Connectivity analyses revealed strengthened coupling between occipital and medial temporal regions, indicating adaptive reconfiguration of the episodic system. The brain does not require images to imagine: it flexibly builds internal experiences using the representational resources available. Summary ParagraphHow humans reconstruct events that are no longer available to the senses is a fundamental but unresolved question. Remembering the past and imagining the future, known as mental time travel, is a defining feature of human cognition, shaping identity and guiding decisions{superscript 1},{superscript 2}. Prevailing theories assume that such constructions depend on visual imagery, with the minds eye reconstructing events on a visuospatial stage3,. Blindness provides a critical test of this assumption. Across extensive behavioral interviews and multimodal neuroimaging, we find that mental time travel remains phenomenologically intact in blindness, but its scaffolding changes fundamentally. Sighted individuals rely on perceptual detail and activate regions specialized for visual scenes, whereas blind individuals, whether blind from birth or later in life, emphasize thoughts and emotions and reorganize occipital cortex for conceptual strategies,. These contrasting strategies map onto distinct neural signatures, revealing a dissociation between perceptual and conceptual routes to episodic simulation. Together, these findings reveal that the brains capacity to construct internal experience rests on conceptual scaffolding, not perceptual re-creation.

Placebo analgesia has traditionally been explained by top-down cortical regulation of brainstem and spinal pathways. Recent circuit-level work in animal models identified a rostral anterior cingulate-pontine-cerebellar pathway that contributes to expectation-based analgesia, implicating cerebellum circuits in placebo effects1. Building on these findings, we examined pontine and cerebellar contributions within a large individual-participant meta-analysis of human neuroimaging studies of placebo analgesia2 (n = 603). We found that the effects of human pain and placebo converge in cerebellar territories embedded in higher-order cognitive3,4 and action-mode networks5. These regions exhibit placebo-induced anticipatory increases and reduced responses during painful stimulation, which correlate with the magnitude of placebo analgesia, consistent with predictive configuration of the system. Pontine responses also correlate with individual differences in placebo analgesia. In independent Human Connectome Project data (n = 820), pontine activity is functionally connected with cingulate and cerebellar regions implicated in placebo analgesia. Together, these findings support a model in which expectation effects are implemented via predictive configuration of a cortico-pontine-cerebellar system.

Chromosomal instability (CIN) arising from mitotic errors generates pervasive structural and numerical chromosome alterations fueling cancer evolution1-3. Entrapment of missegregated chromosomes within micronuclei exacerbates CIN by fostering repeated rounds of aberrant mitotic segregation4-6 and, following micronucleus rupture, promoting catastrophic chromosomal rearrangement processes such as chromothripsis5-9. Whether cellular mechanisms exist to restrain micronucleus-driven CIN has remained unclear. Developing a live-cell chromatin acidification sensor, we tracked micronuclei from genesis through subsequent cell cycles and observed frequent whole-micronucleus capture and acidification via the autophagy pathway. Autophagic targeting is selective for micronuclei with nuclear envelope defects seeded at mitotic exit. Our data indicate that these defects drive progressive loss of chromatin-nuclear envelope tethering, producing a mechanically altered state that is recognised by the autophagic machinery. Autophagy and rupture represent distinct micronuclear fates with opposing genomic consequences. Single-cell sequencing of fate-matched cells demonstrates complete digestion of the chromosomal contents of autophagy-targeted micronuclei, a process we term chromophagy (chromosome-autophagy). By eliminating micronuclei, chromophagy promotes chromosomal loss and arrests the intergenerational transmission of missegregated chromosomal material. This constrains the mutational consequences of micronucleation and suppresses micronucleus-mediated CIN.

Gene expression during mammalian development is orchestrated by non-coding cis-regulatory DNA elements (CREs) such as distal enhancers1-3. Despite their fundamental importance, and notwithstanding recent progress in predictive modeling4-9, many high-level properties of enhancer grammar remain unresolved. How does the length of an autonomously active CRE constrain its activity? How robust are CREs to mutations or rearrangements of transcription factor binding sites (TFBSs)? And how much epistasis exists among these sites? As predictive models solely trained on endogenous CREs are unlikely to resolve these questions10, we subjected several endogenous CREs to intensive sequence-level perturbation. Specifically, we assayed >35,000 variants of 5 parietal endoderm enhancers, with variants organized into four perturbation classes, designed to probe: (i) the functional sufficiency of sub-fragments via dense multi-size tiling, (ii) local epistasis via multi-hit saturation mutagenesis, (iii) activity-size tradeoffs via model-guided compaction, or (iv) functional resilience via sequence derivatization anchored on key TFBSs, including random deposition, reconstitution, and synthetic thripsis. This multi-scale dissection revealed rich phenomena. Sub-tiling uncovered sharp non-additivity between activity and fragment size, highlighting strongly synergistic TFBS clusters. Compaction showed that natural CREs lie far from the activity-size Pareto front, and that model-guided deletions can yield shorter yet stronger elements. Mutational scanning exposed a spectrum of CRE robustness, from tolerant to fragile, together with rare but consequential epistasis between individual TFBSs. Finally, TFBS-anchored derivatization demonstrated that background sequence can influence activity on par with TFBS arrangement. Strikingly, a substantial fraction of CRE derivatives exceeded the activity of their endogenous progenitors. Taken together, these results reveal both soft and stiff directions in regulatory sequence space, advancing a quantitative phenomenology of how enhancer sequences encode function and robustness.

A longstanding challenge in developmental science is to understand how children learn language from naturalistic everyday input. To study this process, we leveraged the First 1,000 Days (1kD) dataset, which provides longitudinal, ultra-dense daily audiovisual recordings for individual children in their home environments. This unusually detailed, child-specific record of early experience enabled us to pair each childs rich language input with a cognitively grounded learning agent, linking naturalistic experience ("nurture") to internal learning mechanisms ("nature"). Trained incrementally on each childs input without prior linguistic knowledge, the learning agent discovered speech units corresponding to the English phoneme inventory and acquired thousands of words, closely mirroring individual developmental trajectories. Learning generalized across children while preserving individual differences in rate and timing. Interestingly, learning relied not only on linguistic input but also on the rehearsal of past experiences at the end of each training day. These findings demonstrate that everyday environments provide sufficient structure for language acquisition and establish a unified mechanistic framework for studying development in real-world contexts.

Archean cratons may provide stable microbial habitats in the deep subsurface, as evidenced by the discovery of billion-year-old crustal fluids1,2. However, the long-term habitability of these cratonic environments is uncertain, as polymetamorphic evolution in most cratons typically destroys microbial habitats through mineral reactions and porosity loss3,4. Preservation of deep microbial habitats is more likely where mantle-derived magma intruded the craton after metamorphic overprinting4. Here we report the discovery of dense microbial colonization at 814 m depth within the 2.05-billion-year-old, unmetamorphosed Bushveld Igneous Complex intrusion, South Africa5. Using advanced contamination-control protocols6,7 and synchrotron-based X-ray spectroscopy, we identified indigenous microbial cells localized at the rims of phlogopite, a hydrous phyllosilicate mineral. Our study reveals that microbial colonization is associated with Fe(III) derived from the structure of phlogopite, where the dehydrogenation likely oxidizes Fe(II) to Fe(III) coupled to H2 generation8. Despite the absence of fracture-driven fluid ingress in the unfractured rock matrix, aqueous alteration evidenced at the rims by potassium removal indicates a self-sustaining habitat driven by an internal redox gradient9. These findings demonstrate that aqueous alteration of ultramafic rocks can sustain isolated microbial life over geological timescales, significantly expanding the potential for long-term habitability on both Earth and Mars4,10.

Defense-associated reverse transcriptases (DRTs) are widespread in bacteria1,2, but how multi-domain DRTs containing RT and additional catalytic activities coordinate antiviral defense remains unclear. Here we show that DRT7, which contains both reverse transcriptase (RT) and primase-polymerase (PP) domains, provides broad-spectrum anti-phage immunity through abortive infection and can be activated by a phage-encoded putative transcriptional regulator. Upon activation, DRT7 synthesizes long, protein-primed, palindromic poly(A)/poly(T)-rich duplex-like DNA. Cryo-electron microscopy structures reveal that RT initiates protein-primed, protein-templated, sequence-specific poly(T) synthesis through an arginine-rich recognition pocket without requiring a complementary nucleic acid template, thereby converting DRT7 from an inactive closed dimer to an active open dimer. The RT-produced poly(T) then serves as both primer and template for PP-mediated poly(A) extension, with iterative handoff between RT and PP generating palindromic, alternating poly(A)/poly(T) ssDNA tracts that assemble into fold-back duplex-like DNA. These findings uncover an unexpected antiviral strategy based on protein self-templating, sequence-specific duplex-like DNA synthesis and reveal how coupling RTs with additional catalytic activities expands the functional scope of nucleic acid synthesis pathways.

Artificial intelligence holds the potential to expand the frontier of scientific research1, yet recent work has raised concern that it may instead narrow scientific attention to well-established areas2-4. Here, leveraging the 2021 release of AlphaFold25 as a quasi-experimental opportunity, we provide field-level evidence that AI can redirect collective attention toward more novel research targets. Tracking 245,396 experimental structures in the Protein Data Bank6, we show that a long-running decline in the study of novel proteins halted after AlphaFold2s release, with the shift concentrated among studies citing AlphaFold2 and targets with high-confidence predictions. This pattern extends to 248,191 downstream papers that consume structural knowledge, where engagement with genes lacking experimental structures and with understudied human genes increased since 2021. Amid rising concern that AI may reinforce scientific canons7-10, our findings offer an early field-level case where AI predictions expand scientific frontiers, consistent with the idea that the real-world consequences of AI on science depend on where their informational gains are greatest. These results suggest AI can complement human knowledge and redirect collective attention in science, with broad implications for emerging AI for science models.

Motivation: Complex disorders arise from multiple genetic mechanisms, but most drug-prioritization methods treat each disorder as a single phenotype and therefore miss locus-specific therapeutic opportunities. Results: We present SIEVE, a framework that decomposes complex disorders into genetically localized subphenotypes and links GWAS summary statistics, reference expression, and perturbational transcriptional profiles to prioritize compounds that target locus-anchored disease mechanisms. SIEVE also constructs genetically calibrated mechanism vectors, projects away nonspecific expression programs using negative anchors, and aggregates evidence across cell lines, doses, and time points to produce robust drug rankings. Across simulations and analyses of real data, SIEVE improves compound prioritization relative to existing methods and shows that subphenotype-aware, genetics-guided modeling can sharpen therapeutic discovery in heterogeneous disorders. Availability and Implementation: R implementation: github.com/ericstrobl/SIEVE.

Sensorimotor associative learning enables animals to adaptively link sensory cues with motor actions, a process that is critical for survival and everyday behavior. While Hebbian mechanisms explain associations formed through temporally overlapping neural activity, a fundamental challenge arises when sensory stimuli and motor responses are separated by a delay, because sensory and motor neurons are rarely coactive. Here, we identify the rostro-lateral posterior parietal cortex (PPC-rl) as a cortical hub that bridges tactile stimuli and temporally delayed licking actions during sensorimotor associative learning. Using cortex-wide calcium imaging with single-cell resolution to track [~]16,000 neurons simultaneously across sensory, motor, and association cortices, we find that PPC-rl uniquely exhibits sustained neural activity during the temporal delay early in learning, a signature that diminishes with expertise. Optogenetic silencing of this activity slows learning without impairing sensorimotor execution in expert mice. Learning strengthens the coupling of population dynamics within and between somatosensory and motor cortices. PPC-rl mediates this process by amplifying a low-dimensional communication subspace that synchronizes co-fluctuations across the somatosensory and motor cortices to facilitate linking. This PPC-rl dependent co-fluctuation dissolves post learning, underscoring PPC-rls role in bridging sensation to distal action. A biologically plausible network indicates that Hebbian plasticity with an eligibility trace gated by reward, PPC-rl persistent activity and PPC-rl dependent sensorimotor subspace communication synergize to support delayed association. Together, our findings uncover a PPC-rl based circuit mechanism that maintains temporal continuity to guide associative learning when sensory and motor events are separated in time.

Single cell genomics has enabled analysis of human prenatal development at unprecedented resolution. However, most studies have relied on dissociated tissues during restricted windows of development, limiting insights into how spatially distributed networks of cells, and multicellular niches emerge and adapt to distinct organ microenvironments in situ. Moreover, existing human developmental atlases have not yet been harmonised, and we thus lack a comprehensive catalogue of known cell types in the developing human body. Here, we introduce the Human Developmental Cell Atlas (HDCA), a unified structural, cellular and molecular resource for prenatal human development. The HDCA integrates published and unpublished single cell/nucleus RNAseq atlases across prenatal organs, and includes a newly generated, spatially resolved, multimodal cell atlas of intact human embryos. Spanning 4-22 post conceptional weeks, capturing embryonic and early to mid fetal stages, the HDCA contains [~]4.6 million cells/nuclei which resolve into [~]450 cell types, explorable with a bespoke web portal. For a global overview of the human embryos multicellular communities, we applied unsupervised deep learning to our intact human embryo spatial data, charting 114 tissue niches that are structural and signalling hubs for the cellular interactions of the embryo. Guided by these niches, we profiled cellular networks over space and time, not examinable using single-organ atlases. In so doing, we revealed tissue-specific fibroblast patterning from previously undescribed mesenchyme progenitors, early diversification of organ-specific blood capillaries and lymphatic vasculature, emergence of neural crest cell fates, the formation of placode-and neural crest-derived peripheral sensory neurons, and how tissue niches guide peripheral neuron maturation and axonal migration. The HDCA thus serves as a comprehensive step towards a comprehensive understanding of human prenatal development, and a template towards unravelling the biology of congenital disorders.

Sleep transforms fragile experiences into lasting memories, but the neuronal basis of this process in humans has remained elusive. In rodents, hippocampal ripples orchestrate the replay of place cell sequences, establishing a cellular mechanism for consolidation - though with limited generalizability to human memory. In humans, neuroimaging has revealed large-scale offline reactivation, but these coarse signals leave open whether individual neurons are reactivated and how ripples might mediate this process. Here, we bridge this gap by directly recording 1,466 medial temporal lobe (MTL) neurons and intracranial electroencephalography during learning, post-learning wakefulness, and sleep. We show that ripples robustly drive neuronal firing, with sleep ripples eliciting stronger activation than wake ripples. Critically, neurons tuned to items that were later remembered fired more strongly during ripples than those coding for forgotten items, and this memory-linked reactivation was selectively observed during sleep. Finally, ripple-associated neuronal MTL bursts were detectable across widespread cortical activity, pointing to a mechanism for systems-level consolidation. Together, these findings provide the first direct evidence that ripple-driven single-neuron reactivation supports human episodic memory consolidation and reveal why sleep -- compared to wakefulness -- offers a privileged window for stabilizing memories.

Understanding and modeling how the human genome encodes gene regulatory programs for thousands of cell types remains a central challenge in genomics and machine learning. However, most human cell types emerge during embryonic, fetal, and pediatric development which are inaccessible to comprehensive molecular profiling. To overcome this, we hypothesized that the mismatch in evolutionary rates between cis-acting enhancers (fast) and the trans-acting regulatory programs that interpret them (slow) creates an opportunity for evolutionary transfer learning. Specifically, models trained to predict cell type-specific enhancers in one species should generalize to the orthologous cell types and enhancers of related species. To test this, we generated a single-cell atlas of chromatin accessibility spanning mouse embryonic day 10 (E10) to birth (P0). Using combinatorial indexing1, we profiled 3.9 million nuclei from 36 staged embryos, resolving genome-wide accessibility in 36 cell classes and 140 cell types. With the goal of identifying distal enhancers for all cell classes, we trained a series of multi-output deep learning models (CREsted2), each addressing limitations of the preceding approach. An evolution-naive model achieves strong performance on heldout peaks, but exhibited two failure modes during genome-wide inference: overprediction at tandem repeats and conflation of promoter and distal enhancer grammars. An evolution-aware model resolves these by regrouping accessible regions based on functional coherence across syntenic orthologs, but fails to generalize across species -- suggesting insufficient sequence diversity during training. Finally, STEAM (Synteny-aware Transfer learning for Enhancer Activity Modeling), our evolution-augmented model, expands the training corpus to include enhancer orthologs from up to 241 mammalian genomes (Zoonomia3) in a synteny-supervised manner. This increases the effective data scale by up to 195-fold, markedly improving generalization across mammals despite greater label noise. We apply STEAM predict enhancers for all major developmental lineages throughout the human, mouse (HumMus) and 239 additional mammalian genomes3 (BabaGanoush), i.e. 32 x 241 = 7,712 genome-wide enhancer tracks. Together, our results unify advances in single-cell profiling, deep learning, and comparative genomics into a framework for the evolutionary transfer learning of noncoding regulatory grammars. More broadly, our work supports the view that model organisms and evolutionarily diverse genomes are indispensable resources for accelerating the AI-enabled exploration of human biology. NoteAn interactive version of this preprint, together with count matrices, CREsted models, prediction tracks, code and reproducible figures, is available at this link (ref 4).

Bacteria have evolved diverse immune strategies to detect and neutralize bacteriophage infection. Here, we describe an unprecedented paradigm in which a chromosome-architecting nucleoid-associated protein (NAP) is repurposed as a viral infection sensor. When phage attack leads to genome degradation, the NAP sensor is released from the nucleoid to the cytoplasm, where it binds and activates diverse immune effectors. One such effector is a nucleotide-modifying toxin normally existing as an inactive homotetramer. NAP binding converts it into a catalytically active heterotrimer that halts both transcription and translation. Phylogenetic analyses unveiled the high modularity, polyphyletic origin, and wide distribution of NAP-mediated defenses. Collectively, we define a distinct class of defense systems in which bacteria sense phage-induced genome damage through NAP relocation, highlighting an unexpected but essential role for these proteins as sentinels of genome integrity.

Animals rely on olfaction to locate food, mates, and suitable habitats, yet natural odour environments contain thousands of volatile molecules, creating a high-dimensional sensory problem for both nervous systems and the researchers who study them 1-5. For example, a banana emits around 100 individual volatiles4,6. It remains unclear which components of complex odour blends animals have evolved to use as behavioural cues. Here, combining fieldwork, chemical and behavioural analyses, we show across multiple Drosophila species that behaviourally relevant cues can be predicted directly from the statistical structure of natural odour environments. Animals preferentially respond to components that are most distinctive within their natural host odour blends, and therefore most ecologically informative. These cues can be either major or minor blend components. Our results indicate that host-guided olfactory behaviours have evolved to exploit the statistical structure of natural odour environments by selectively targeting the most informative features of odour blends.

AO_SCPLOWBSTRACTC_SCPLOWThe biology that governs progression and therapeutic response in autoimmune disease is organized in affected tissue, but direct molecular readout of that biology requires invasive biopsy and is rarely repeated during clinical trials or routine care. Using paired blood-skin single-cell RNA-sequencing from a systemic sclerosis (SSc) cohort of 74 individuals (57 patients and 17 matched controls, 192,809 cells across 53 annotated cell states), we show that peripheral blood carries a recoverable projection of tissue-resident molecular state. Across 63 pathways scored in both compartments, 43 same-pathway blood-skin associations reach FDR < 0.05; at cell-type resolution, 212 cross-compartment associations survive residualization for disease status and sex. Per-patient classifiers recover tissue-defined molecular states out of fold with AUCs between 0.62 and 0.79, with the strongest recoveries on fibroblast subtype programs that have no direct circulating analog: fibroblast COMP at 0.79, COCH at 0.75, MYOC2 at 0.74, POSTN at 0.74. Tissue programs route through different blood compartments at different representational levels: fibroblast programs resolve through T-cell, Treg, monocyte and B-cell axes at compositional and distributional levels, while interferon resolves through expression state across multiple cell types. Within SSc alone, a cross-validated partial least squares model learns a shared blood-skin latent axis at r = 0.486 (permutation p = 0.006); the induced patient ranking recovers tissue-interferon-high patients at 86% precision at the top-20% screening threshold against a 50% base rate. A paired multiview autoencoder, trained on module-level dependency structure under contrastive alignment, paired reconstruction, neighborhood preservation and tissue-target supervision, learns a shared latent geometry in which blood-only projections land in the same tissue-state region as their matched tissue samples and supports recovery of held-out tissue targets above simpler baselines and above two permutation null families. These results map the empirical geometry of cross-compartment inference in autoimmune disease and position peripheral blood as a substrate for tissue-state inference at trial and clinical scale.

Crossovers are essential for accurate chromosome segregation in meiosis. Yet the programmed DNA double-strand breaks that initiate them frequently occur in genes and pose a risk to transcription required for gametogenesis. How meiotic cells reconcile these competing demands has remained unclear. Here, we generate a genome-wide in vivo atlas of meiotic recombination intermediates across [~]42,000 hotspots by mapping repair proteins BLM, HFM1, and RPA in wild-type and genome-engineered mutant mouse testes. These maps reveal two distinct modes of break repair: a fast-resolving class with short-lived intermediates that are repaired predominantly as non-crossovers, and a slower class with persistent intermediates that give rise to nearly all crossovers. Fast-resolving hotspots occur almost exclusively within a deeply conserved set of [~]4,500 genes marked by structural and chromatin features established during an early stage of meiotic transcription. This transcriptional memory predicts repair fate with high accuracy across mouse subspecies and sexes. Across widely diverged mammals, including humans and cattle, orthologous genes show similar crossover suppression. Our findings reveal an early bifurcation between crossover and non-crossover repair that is governed by the transcriptional context of meiotic breaks. Together, they establish an evolutionarily conserved principle in which crossovers are directed away from transcriptionally important genes, thereby safeguarding gene function and shaping their evolution.