Understanding the selective forces acting upon HIV early in infection is crucial to design prevention strategies. By leveraging deep sequencing and the short diagnostic intervals of the FRESH and RV217 cohorts (median 4 days) between the last-negative and first-positive RNA tests, we captured a precise and early snapshot of acute HIV infection. The frequency of multiple transmitted viruses of 38% in these as well as placebo recipients from the AMP trials was higher than previously published, with the true frequency likely to be higher. The relative abundance of lineages fluctuated substantially over time in two-thirds of the multilineage infections, generating uncertainty in identifying the specific viruses that were transmitted and founding the infection. Viral populations exhibited diversity and selection on the Gag and Env proteins at the earliest times examined, with sites inferred to be undergoing negative selection most evident. These data may help explain vaccination failures and provide new targets for prevention.

Abstract In the antibody mediated prevention (AMP) trials, the broadly neutralizing antibody (bNAb) VRC01 demonstrated protective efficacy against susceptible HIV strains. To understand how VRC01 shaped breakthrough infections, deep sequencing was performed on 172 participants (>100,000 gag-Δpol and rev-env-Δnef sequences), at diagnosis and over time, in the placebo and treatment arms of the African (HVTN703/HPTN081; NCT02568215) and Americas/Europe (HVTN704/HPTN085; NCT02716675) cohorts. A high frequency of multilineage infections was detected (38%), including co-infection with both VRC01 sensitive and resistant viruses. This high frequency is largely accounted for by low-abundance lineages. Although VRC01 does not significantly affect the genetic transmission bottleneck compared to placebo, higher VRC01 doses trend towards greater VRC01 neutralization differences among co-infecting lineages. Two-thirds of multilineage infections showed evidence of recombination at the diagnostic timepoint. In the treatment group there is evidence of recombinant viruses preferentially inheriting resistance-associated mutations. This study provides critical insights into viral genetic and antigenic diversity that needs to be targeted to achieve protection, and highlights the role of recombination in facilitating escape.

Standard probabilistic models of coding sequence evolution effectively identify where and when selection acts but remain agnostic to the mechanistic realization of these forces. We introduce PRIME (PRoperty Informed Models of Evolution), a framework of codon-level maximum likelihood methods-including global (G-PRIME), episodic (E-PRIME), and site-specific (S-PRIME) implementations-that explicitly model amino acid exchangeability as a function of physicochemical properties. By parameterizing attributes such as molecular volume, hydropathy, and secondary structure propensities, PRIME resolves the biophysical basis of selective constraint across both the sequence and the phylogeny. At the site level, S-PRIME leverages an explicit biophysical taxonomy to precisely categorize residues as conserved, neutral, or changing for specific properties, resolving selective signals that remain invisible to traditional rate-based metrics. Our analysis of a benchmark of 24 diverse datasets and a genome-wide screen of 18,944 mammalian genes demonstrates that biophysical realism yields substantial improvements in model fit, acting synergistically with rate variation to explain complex evolutionary patterns. We find that power to detect physicochemical constraints at individual sites is fundamentally governed by simple informational redundancy (substitutions per unique amino acid; AUC = 0.91), with sensitivity exceeding 90% in data-rich alignments. E-PRIME reveals a distinct biophysical hierarchy: while core packing and beta-sheet scaffolds are rigidly conserved, alpha-helix propensity and surface electrostatics serve as the primary substrates for adaptive tuning. Furthermore, PRIME importance weights align with aspects of the primary semantic axes of deep learning representations (ESM-2) and capture key features of experimental fitness landscapes. By transforming abstract evolutionary rates into interpretable biophysical rules, PRIME provides a useful framework for characterizing the mechanistic drivers of protein diversity.

Standard probabilistic models of coding sequence evolution effectively identify where and when selection acts but remain agnostic to the mechanistic realization of these forces. We introduce PRIME (PRoperty Informed Models of Evolution), a framework of codon-level maximum likelihood methods-including global (G-PRIME), episodic (E-PRIME), and site-specific (S-PRIME) implementations-that explicitly model amino acid exchangeability as a function of physicochemical properties. By parameterizing attributes such as molecular volume, hydropathy, and secondary structure propensities, PRIME aims to resolve the biophysical basis of selective constraint across both the sequence and the phylogeny. At the site level, S-PRIME leverages an explicit biophysical taxonomy to categorize residues as conserved, neutral, or changing for specific properties, resolving selective signals that are missed by traditional rate-based metrics. Our analysis of a benchmark of 24 diverse datasets and a genome-wide screen of 18,944 mammalian genes demonstrates that consideration of biophysical realism can yield substantial improvements in model fit, acting synergistically with rate variation to explain complex evolutionary patterns. We find that physicochemical constraints at individual sites can be reliably detected in datasets with sufficient information redundancy (substitutions per unique amino acid; AUC=0.91), with sensitivity exceeding 90% in data-rich alignments. E-PRIME reveals a distinct hierarchy in biophysical constraints: while core packing and beta-sheet scaffolds are rigidly conserved, alpha-helix propensity and surface electrostatics serve as the primary substrates for adaptive tuning. Furthermore, PRIME importance weights align with aspects of the primary semantic axes of deep learning representations (ESM-2) and capture key features of experimental fitness landscapes. By transforming abstract evolutionary rates into interpretable biophysical rules, PRIME provides a useful framework for characterizing the mechanistic drivers of protein diversity.

Population-based studies of circulating blood proteins have provided profound insights into human biology. However, short-lived changes often remained undetected. To address this, we performed a comprehensive longitudinal dried blood spot (DBS) self-sampling study in 808 young adults of the BAMSE cohort during 2020-2022. We profiled serological, autoimmune, and proteomic phenotypes in relation to SARS-CoV-2 exposures (infection, vaccination), physiological traits, genetic variation, and blood counts. Data-driven seroclustering revealed dynamic immune response to both exposures, while analysis of anti-interferon autoantibodies (AAbs) uncovered associations of pre-existing stable AAbs with prolonged COVID-19 symptoms. Genome-wide mapping determined 664 pQTLs, with cis-pQTLs associated showing increased longitudinal stability. We also identified relationships between blood cell counts and DBS proteins beyond the hematocrit effects. Paired pre- and post-exposure highlighted transient alterations for infection (e.g. LAP3) and vaccinations (e.g. TIMP3). Multi-molecular phenotyping in self-sampling can capture dynamic immune trajectories, informing precision medicine efforts with clinically valuable insights on short-term variability in health phenotypes. ### Competing Interest Statement NR is a co-founder and shareholder of the microsampling companies Capitainer AB and Samplimy Medical AB, and an inventor of several patents on microsampling solutions. JMS is a Scientific Advisor for ABC Labs and has, unrelated to this work, received travel support from Olink, Alamar Bioscience, Illumina, Oxford Nanopore and Luminex, and via KTH, conducted contract research for Capitainer and Luminex. All other authors declare no competing interests. ### Funding Statement The BAMSE study received supported from grants of the Swedish Research Council (2016‐03086; 2018‐02524; 2019‐01060; 2020‐02170; 2022-06340; 2024-03164); Forskningsrådet för hälsa, Arbetsliv och välfärd (2017‐00526); Formas (2016‐01646); Hjärt‐Lungfonden; Region Stockholm (ALF project); and the Asthma and Allergy Research Foundation. JMS received funding from the SciLifeLab National COVID-19 Research Program, which was financed by the Knut and Alice Wallenberg Foundation (2020.0182, 2020.0241) and SciLifeLab's Pandemic Laboratory Preparedness program (VC-2022-0028). NR received funds from Sweden's innovation agency Vinnova (2020-04451). BM received funds form The Erling Persson Foundation (20210125). We also acknowledge the tremendous support from the Knut and Alice Wallenberg Foundation for funding the Human Protein Atlas. This work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Informed written consent was obtained from all subjects, and ethical permits were granted by the Regional Ethical Review Board in Stockholm (2016/1380-31/2) and Swedish Ethical Review Authority (2020-02922, 2024-06052-02). I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes The datasets of the BAMSE study are not publicly available due to legal and ethical regulations. Upon reasonable requests, data supporting the findings of this study are available from the PI of the BAMSE cohort (Professor Erik Melén, erik.melen{at}ki.se).

Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) has emerged as a central approach for studying T cell and B cell receptor populations, and is now an important component of studies of autoimmunity, immune responses to pathogens, vaccines, allergens, and cancers, and for antibody discovery. When amplifying the rearranged V(D)J genes encoding antigen receptors, each cycle of the Polymerase Chain Reaction (PCR) can produce spurious "chimeric" hybrids of two or more different template sequences. While the generation of chimeras is well understood in bacterial and viral sequencing, and there are dedicated tools to detect such sequences in bacterial and viral datasets, this is not the case for AIRR-seq. Further, the process that results in immune receptor sequences has domain-specific challenges, such as somatic hypermutation (SHM), and domain-specific opportunities, such as relatively well-known germline gene "reference" sequences. Here we describe CHMMAIRRa, a hidden Markov model for detecting chimeric sequences in AIRR-seq data, that specifically models SHM and incorporates germline reference sequences. We use simulations to characterize the performance of CHMMAIRRa and compare it to existing methods from other domains, we test the effect of PCR conditions on chimerism using IgM libraries generated in this study, and we apply CHMMAIRRa to four published AIRR-seq datasets to show the extent and impact of artifactual chimerism.

Generative models are becoming powerful tools for protein design, enabling the creation of novel protein structures and sequences. Recent approaches have shown success using diffusion models and flow matching to sample realistic protein folds. Many methods explicitly incorporate structural biases or constraints—for example, enforcing symmetry during generation—to steer the design process. Here we report the spontaneous emergence of structural symmetry in a transformer-based generative model for proteins, without any symmetry-specific conditioning during training or constraints during generation. In our flow-matching model, a single attention head in an SE(3)-equivariant transformer layer was found to be primarily responsible for the model’s ability to generate symmetric arrangements of backbone residues across chains, or repeating motifs within a chain. Our results show that protein generative models can learn high-level structural patterns implicitly from training data. This opens new questions about interpretability and control in generative design: understanding how and why a single attention head can govern a complex global property like symmetry may inform future model architectures and help exploit emergent behaviors for better protein engineering. ### Competing Interest Statement The authors have declared no competing interest. Swedish Research Council, https://ror.org/03zttf063, 2023-02516

The COVID-19 pandemic showcased a coevolutionary race between the human immune system and SARS-CoV-2, during which the immune system generated neutralizing antibodies targeting the SARS-CoV-2 spike protein's receptor-binding domain (RBD), crucial for host cell invasion, while the virus evolved to evade antibody recognition. Here, we establish a synthetic coevolution system combining high-throughput screening of antibody and RBD variant libraries with protein mutagenesis, surface display, and deep sequencing. Additionally, to significantly extend our interrogation of sequence space, we train a protein language model that predicts antibody escape to RBD variants and demonstrate its capability to generalize to a larger mutational load and mutations at positions unseen during training. Through explainable AI techniques, we probe the model and identify biologically meaningful coevolution trends. Synthetic coevolution reveals antagonistic and compensatory mutational trajectories of neutralizing antibodies and SARS-CoV-2 variants, enhancing the understanding of this evolutionary conflict.

While many phylogenetic methods exist to characterize evolutionary pressure at individual codon sites, relatively few allow direct comparison between different a priori selected sets of branches. Such comparisons may be useful for pinpointing precisely the codon sites that are under differing selective pressures due to differing environmental contexts, or differing genomic contexts via epistatic interactions. Indeed, this was only recently addressed by an approach, developed in the frequentist framework, that proposes a site-wise likelihood ratio hypothesis test. Previously, we have demonstrated that approximate grid-based Bayesian approaches to characterizing site-wise variation in selection parameters can outperform individual site-wise likelihood ratio tests. Such grid-based approaches can exhibit poor computational scaling when the number of site-wise parameters expands, but here we show that this is still tractable up to four parameters, and that a simple subtree-likelihood caching strategy can provide efficiency improvements in some cases. We propose difFUBAR, which allows the demarcation of two branch sets of interest and, optionally, a background set, and estimates joint site-specific posterior distributions over α, ω 1, ω 2, and ω BG using a Gibbs sampler. Evidence for hypotheses of interest can then be quantified directly from the posterior distribution, and we standardly report P ( ω 1 > ω 2), P ( ω 2 > ω 1), P ( ω 1 > 1), and P ( ω 2 > 1). We characterize the computational and statistical performance of this approach on previous simulations, comparing it to the site-wise likelihood ratio test approach, where it shows moderate statistical benefits, and substantial computational gains, typically being more than two orders of magnitude faster on the same datasets. We also demonstrate that it can scale to datasets of over ten thousand taxa, on a laptop in under ten minutes. difFUBAR is implemented in MolecularEvolution.jl - a Julia framework for phylogenetic model development - and can be run locally, or online via a Colab notebook. ### Competing Interest Statement The authors have declared no competing interest. Swedish Research Council, https://ror.org/03zttf063, 2022-05034

Protective antibodies against HIV-1 require unusually high levels of somatic mutations introduced in germinal centers (GCs). To achieve this, a sequential vaccination approach was proposed. Using HIV-1 antibody knockin mice with fate-mapping genes, we examined if antigen affinity affects the outcome of B cell recall responses. Compared to a high-affinity boost, a low-affinity boost resulted in decreased numbers of memory-derived B cells in secondary GCs but with higher average levels of somatic mutations, indicating an afcomposition of primary GCs was modified in an antigen-affinity-dependent manner to constitute less somatically mutated B cells. Our results demonstrate that antigen affinity and location of the boost affect the outcome of the B cell recall response. These results can help guide the design of vaccine immunogens aiming to selectively engage specific B cell clones for further diversification.

Broadly neutralizing antibodies (bnAbs) show promise in HIV prevention, yet viral escape remains a challenge. In the Antibody Mediated Prevention (AMP) trials, the CD4 binding site (CD4bs) bNAb VRC01 blocked acquisition by VRC01-sensitive strains. However, its influence on viral evolution post-acquisition is not fully understood. Here we analyzed >12,000 HIV env sequences from 47 participants from the AMP trials, identifying VRC01-mediated de novo escape mutations in 8 of 26 VRC01-treated participants but none in 21 placebo participants. These mutations were found at very low frequency (<1%) in global viruses. Escape mutations, primarily located in the Loop-D and β23/V5 regions of Env, conferred cross-resistance to several CD4bs bnAbs, while more potent CD4bs bnAbs like N6 and 1-18 largely retained their activity. Our findings demonstrate that prophylactic VRC01 can select for viral escape after infection, underscoring the need for next-generation bnAbs with improved breadth and potency to enhance durability and efficacy of antibody-based HIV prevention.

MOTIVATION:Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) has emerged as a central approach for studying T cell and B cell receptor populations, and is now an important component of studies of autoimmunity, immune responses to pathogens, vaccines, allergens, and cancers, and for antibody discovery. When amplifying the rearranged V(D)J genes encoding antigen receptors, each cycle of the Polymerase Chain Reaction (PCR) can produce spurious "chimeric" hybrids of two or more different template sequences. While the generation of chimeras is well understood in bacterial and viral sequencing, and there are dedicated tools to detect such sequences in bacterial and viral datasets, this is not the case for AIRR-seq. Further, the process that results in immune receptor sequences has domain-specific challenges, such as somatic hypermutation (SHM), and domain-specific opportunities, such as relatively well-known germline gene "reference" sequences. RESULTS:Here, we describe CHMMAIRRa, a hidden Markov model for detecting chimeric sequences in AIRR-seq data, that specifically models SHM and incorporates germline reference sequences. We use simulations to characterize the performance of CHMMAIRRa and compare it to existing methods from other domains, we test the effect of PCR conditions on chimerism using IgM libraries generated in this study, and we apply CHMMAIRRa to four published AIRR-seq datasets to show the extent and impact of artifactual chimerism. AVAILABILITY AND IMPLEMENTATION:CHMMAIRRa is published on the Julia package registry and is available at https://github.com/MurrellGroup/CHMMAIRRa.jl (DOI: 10.5281/zenodo.17279881). The core HMM implementation is available at https://github.com/MurrellGroup/CHMMera.jl (DOI: 10.5281/zenodo.17279998), and the scripts used to generate the results in this paper at https://github.com/MurrellGroup/CHMMAIRRaAnalyses (DOI: 10.5281/zenodo.17281446).

Just as language is composed of sublexical tokens that combine to form words, sentences, and paragraphs, protein backbones are composed of sub-structural elements that combine to form helices, sheets, folds, domains, and chains. Autoregressive language models operate on discrete tokens, whereas protein structure is inherently continuous, and generative approaches to protein design have borrowed more from image generation than language modeling. But autoregressive models do not inherently require their inputs and outputs to be discrete. Here we describe a generative autoregressive language model over the continuous space of protein backbones, where the distribution over the placement of each successive amino acid is conditioned on all preceding residues, and can be sampled from one residue after another. We show that this approach can learn to sample diverse and realistic protein chains, opening a new potential avenue for in silico protein design.### Competing Interest StatementThe authors have declared no competing interest.

Descendants of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron variant now account for almost all SARS-CoV-2 infections. The Omicron variant and its sublineages have spike glycoproteins that are highly diverged from the pandemic founder and first -generation vaccine strain, resulting in significant evasion from monoclonal antibody therapeutics and vaccines. Understanding how commonly elicited antibodies can broaden to cross -neutralize escape variants is crucial. We isolate IGHV3-53, using "public"monoclonal antibodies (mAbs) from an individual 7 months post infection with the ancestral virus and identify antibodies that exhibit potent and broad cross -neutralization, extending to the BA.1, BA.2, and BA.4/BA.5 sublineages of Omicron. Deep mutational scanning reveals these mAbs' high resistance to viral escape. Structural analysis via cryoelectron microscopy of a representative broadly neutralizing antibody, CAB -A17, in complex with the Omicron BA.1 spike highlights the structural underpinnings of this broad neutralization. By reintroducing somatic hypermutations into a germline-reverted CAB -A17, we delineate the role of affinity maturation in the development of cross -neutralization by a public class of antibodies.

AbstractObjectivesThe caecum bridges the small and large intestine and plays a front‐line role in discriminating gastrointestinal antigens. Although dysregulated in acute and chronic conditions, the tissue is often overlooked immunologically.MethodsTo address this issue, we applied single‐cell transcriptomic‐V(D)J sequencing to FACS‐isolated CD45+ caecal patch/lamina propria leukocytes from a healthy (5‐year‐old) female rhesus macaque ex vivo and coupled these data to VDJ deep sequencing reads from haematopoietic tissues.ResultsWe found caecal NK cells and ILC3s to co‐exist with a spectrum of effector T cells partially derived from SOX4+ recent thymic emigrants. Tolerogenic Vγ8Vδ1‐T cells, plastic CD4+ T helper cells and GZMK+EOMES+ and TMIGD2+ tissue‐resident memory CD8+ T cells were present and differed metabolically. An IL13+GATA3+ Th2 subset expressing eicosanoid pathway enzymes was accompanied by IL1RL1+GATA3+ regulatory T cells and a minor proportion of IgE+ plasma cells (PCs), illustrating tightly regulated type 2 immunity devoid of ILC2s. In terms of B lymphocyte lineages, caecal patch antigen‐presenting memory B cells sat alongside germinal centre cells undergoing somatic hypermutation and differentiation into IGF1+ PCs. Prototypic gene expression signatures decreased across PC clusters, and notably, expanded IgA clonotypes could be traced in VDJ deep sequencing reads from additional compartments, including the bone marrow, supporting that these cells contribute a steady stream of systemic antibodies.ConclusionsThe data advance our understanding of caecal immunological function, revealing processes involved in barrier maintenance and molecular networks relevant to disease.

SARS-CoV-2 variants acquire mutations in spike that promote immune evasion and impact other properties that contribute to viral fitness such as ACE2 receptor binding and cell entry. Knowledge of how mutations affect these spike phenotypes can provide insight into the current and potential future evolution of the virus. Here we use pseudovirus deep mutational scanning to measure how >9,000 mutations across the full XBB.1.5 and BA.2 spikes affect ACE2 binding, cell entry, or escape from human sera. We find that mutations outside the receptor-binding domain (RBD) have meaningfully impacted ACE2 binding during SARS-CoV-2 evolution. We also measure how mutations to the XBB.1.5 spike affect neutralization by serum from individuals who recently had SARS-CoV-2 infections. The strongest serum escape mutations are in the RBD at sites 357, 420, 440, 456, and 473-however, the antigenic impacts of these mutations vary across individuals. We also identify strong escape mutations outside the RBD; however many of them decrease ACE2 binding, suggesting they act by modulating RBD conformation. Notably, the growth rates of human SARS-CoV-2 clades can be explained in substantial part by the measured effects of mutations on spike phenotypes, suggesting our data could enable better prediction of viral evolution.

Against the backdrop of the rapid global takeover and dominance of BA.1/BA.2 and subsequently BA.2.86 lineages, the emergence of a highly divergent SARS-CoV-2 variant warrants characterization and close monitoring. Recently, another such BA.2 descendent, designated BA.2.87.1, was detected in South Africa. Here, we show using spike-pseudotyped viruses that BA.2.87.1 is less resistant to neutralisation by prevailing antibody responses in Sweden than other currently circulating variants such as JN.1. Further we show that a monovalent XBB.1.5-adapted booster enhanced neutralising antibody titers to BA.2.87.1 by almost 4-fold. While BA.2.87.1 may not outcompete other currently-circulating lineages, the repeated emergence and transmission of highly diverged variants suggests that another large antigenic shift, similar to the replacement by Omicron, may be likely in the future. ### Competing Interest Statement DJS has served as consultant for AstraZeneca AB.

The continued evolution of SARS-CoV-2 underscores the need to understand qualitative aspects of the humoral immune response elicited by spike immunization. Here, we combine monoclonal antibody (mAb) isolation with deep B cell receptor (BCR) repertoire sequencing of rhesus macaques immunized with prefusion-stabilized spike glycoprotein. Longitudinal tracing of spike-sorted B cell lineages in multiple immune compartments demonstrates increasing somatic hypermutation and broad dissemination of vaccine-elicited B cells in draining and non-draining lymphoid compartments, including the bone marrow, spleen and, most notably, periaortic lymph nodes. Phylogenetic analysis of spike-specific monoclonal antibody lineages identified through deep repertoire sequencing delineates extensive intra-clonal diversification that shaped neutralizing activity. Structural analysis of the spike in complex with a broadly neutralizing mAb provides a molecular basis for the observed differences in neutralization breadth between clonally related antibodies. Our findings highlight that immunization leads to extensive intra-clonal B cell evolution where members of the same lineage can both retain the original epitope specificity and evolve to recognize additional spike variants not previously encountered.

The immune responses to Novavax’s licensed NVX-CoV2373 nanoparticle Spike protein vaccine against SARS-CoV-2 remain incompletely understood. Here, we show in rhesus macaques that immunization with Matrix-MTM adjuvanted vaccines predominantly elicits immune events in local tissues with little spillover to the periphery. A third dose of an updated vaccine based on the Gamma (P.1) variant 7 months after two immunizations with licensed NVX-CoV2373 resulted in significant enhancement of anti-spike antibody titers and antibody breadth including neutralization of forward drift Omicron variants. The third immunization expanded the Spike-specific memory B cell pool, induced significant somatic hypermutation, and increased serum antibody avidity, indicating considerable affinity maturation. Seven months after immunization, vaccinated animals controlled infection by either WA-1 or P.1 strain, mediated by rapid anamnestic antibody and T cell responses in the lungs. In conclusion, a third immunization with an adjuvanted, low-dose recombinant protein vaccine significantly improved the quality of B cell responses, enhanced antibody breadth, and provided durable protection against SARS-CoV-2 challenge.