Domain Fusion Research Articles

Characterization of a MHYT domain-coupled transcriptional regulator that responds to carbon monoxide.

The MHYT domain, identified over two decades ago for its potential to detect diatomic gases like CO, O2or NO, has awaited experimental validation as a protein sensory domain. Here, we characterize the MHYT domain-containing transcriptional regulator CoxC, which governs the expression of the cox genes responsible for aerobic CO oxidation in the carboxidotrophic bacterium Afipia carboxidovorans OM5. The C-terminal LytTR-type DNA-binding domain of CoxC binds to an operator region consisting of three direct repeats sequences overlapping the -35 box at the target PcoxB promoter, which is consistent with the role of CoxC as a specific transcriptional repressor of the cox genes. Notably, the N-terminal transmembrane MHYT domain endows CoxC with the ability to sense CO as an effector molecule, as demonstrated by the relief of CoxC-mediated repression and binding to the PcoxB promoter upon CO exposure. Furthermore, copper serves as the essential divalent cation for the interaction of CO with CoxC, thereby confirming previous hypothesis regarding the role of copper in the gas-sensing mechanism of MHYT domains. CoxC represents the prototype of a novel subfamily of single-component LytTR transcriptional regulators, characterized by the fusion of a DNA-binding domain with a membrane-bound MHYT sensor domain.

Cascading Failure in Cyber-Physical Systems: A Review on Failure Modeling and Vulnerability Analysis.

Cascading failures pose a significant security threat to networked systems, with recent global incidents underscoring their destructive potential. The security threat of cascading failures has always existed, but the evolution of cyber-physical systems (CPSs) has introduced novel dimensions to cascading failures, intensifying their threats owing to the intricate fusion of cyber and physical domains. Addressing these threats requires a nuanced understanding achieved through failure modeling and vulnerability analysis. By analyzing the historical failures in different CPSs, the cascading failure in CPSs is comprehensively defined as a complicated propagation process in coupled cyber and physical systems, initialized by natural accidents or human interference, which exhibits a progressive evolution within the networked structure and ultimately results in unexpected large-scale systemic failures. Subsequently, this study advances the development of instructions for modeling cascading failures and conducting vulnerability analyses within CPSs. The examination also delves into the core challenges inherent in these methodologies. Moreover, a comprehensive survey and classification of extant research methodologies and solutions are undertaken, accompanied by a concise evaluation of their advancements and limitations. To validate the performance of these methodologies, numerical experiments are conducted to ascertain their distinct features. In conclusion, this article advocates for future research initiatives, particularly emphasizing the exploration of uncertainty analysis, defense strategies, and verification platforms. By addressing these areas, the resilience of CPSs against cascading failures can be significantly enhanced.

An Immunoreceptor-Targeting Strategy with Minimalistic C3b Peptide Fusion Enhances SARS-CoV-2 RBD mRNA Vaccine Immunogenicity.

The clinical success of mRNA vaccine during the COVID-19 pandemic has inspired emerging approaches to elevate mRNA vaccine immunogenicity. Among them, antigen fusion protein designs for improved immune cell targeting have been shown to augment humoral immunity against small antigen targets. This research demonstrates that SARS-CoV-2 receptor binding domain (RBD) fusion with a minimalistic peptide segment of complement component 3b (C3b, residues 727-767) ligand can improve mRNA vaccine immunogenicity through antigen targeting to complement receptor 1 (CR1). We affirm vaccines' antigenicity and targeting ability towards specific receptors through Western blot and immunofluorescence assay. Furthermore, mice immunization studies help the investigation of the antibody responses. Using SARS-CoV-2 Omicron RBD antigen, we compare mRNA vaccine formulations expressing RBD fusion protein with mouse C3b peptide (RBD-mC3), RBD fusion protein with mouse Fc (RBD-Fc), and wild-type RBD. Our results confirm the proper antigenicity and normal functionality of RBD-mC3. Upon validating comparable antigen expression by the different vaccine formulations, receptor-targeting capability of the fusion antigens is further confirmed. In mouse immunization studies, we show that while both RBD-mC3 and RBD-Fc elevate vaccine immunogenicity, RBD-mC3 leads to more sustained RBD-specific titers over the RBD-Fc design, presumably due to reduced antigenic diversion by the minimalistic targeting ligand. The study demonstrates a novel C3b-based antigen design strategy for immune cell targeting and mRNA vaccine enhancement.

A dual domain multi-exposure image fusion network based on spatial-frequency integration

Multi-exposure image fusion aims to generate a single high-dynamic image by integrating images with different exposures. Existing deep learning-based multi-exposure image fusion methods primarily focus on spatial domain fusion, neglecting the global modeling ability of the frequency domain. To effectively leverage the global illumination modeling ability of the frequency domain, we propose a novelty perspective on multi-exposure image fusion via the Spatial-Frequency Integration Framework, named MEF-SFI. Initially, we revisit the properties of the Fourier transform on the 2D image, and verify the feasibility of multi-exposure image fusion in the frequency domain where the amplitude and phase component is able to guide the integration of the illumination information. Subsequently, we present the deep Fourier-based multi-exposure image fusion framework, which consists of a spatial path and frequency path for local and global modeling separately. Specifically, we introduce a Spatial-Frequency Fusion Module to facilitate efficient interaction between dual domains and capture complementary information from input images with different exposures. Finally, we combine a dual domain loss function to ensure the retention of complementary information in both the spatial and frequency domains. Extensive experiments on the PQA-MEF dataset demonstrate that our method achieves visual-appealing fusion results against state-of-the-art multi-exposure image fusion approaches. Our code is available at https://github.com/SSyangguang/MEF-freq.

Engineering psychrophilic polymerase for nanopore long-read sequencing.

Unveiling the potential application of psychrophilic polymerases as candidates for polymerase-nanopore long-read sequencing presents a departure from conventional choices such as thermophilic Bacillus stearothermophilus (Bst) renowned for its limitation in temperature and mesophilic Bacillus subtilis phage (phi29) polymerases for limitations in strong exonuclease activity and weak salt tolerance. Exploiting the PB-Bst fusion DNA polymerases from Psychrobacillus (PB) and Bacillus stearothermophilus (Bst), our structural and biochemical analysis reveal a remarkable enhancement in salt tolerance and a concurrent reduction in exonuclease activity, achieved through targeted substitution of a pivotal functional domain. The sulfolobus 7-kDa protein (Sso7d) emerges as a standout fusion domain, imparting significant improvements in PB-Bst processivity. Notably, this study elucidates additional functional sites regulating exonuclease activity (Asp43 and Glu45) and processivity using artificial nucleotides (Glu266, Gln283, Leu334, Glu335, Ser426, and Asp430). By disclosing the intricate dynamics in exonuclease activity, strand displacement, and artificial nucleotide-based processivity at specific functional sites, our findings not only advance the fundamental understanding of psychrophilic polymerases but also provide novel insights into polymerase engineering.

GRU-AE-wiener: A generative adversarial network assisted hybrid gated recurrent unit with Wiener model for bearing remaining useful life estimation

Bearings play a pivotal role in various mechanical systems, and their health directly impacts the reliability and safety of these systems. Consequently, extensive research has been dedicated to the estimation of Bearing Remaining Useful Life (RUL) through the lens of information fusion theory. The absence of comprehensive life-cycle degradation data for bearings, a common challenge within the information fusion domain, can hinder the accuracy and reliability of RUL prediction models. A novel hybrid data and model approach named GRU-AE-Wiener has been developed to address this limitation. This approach combines the power of Gated Recurrent Unit (GRU) and Wiener process models within the information fusion framework. Firstly, a Loop Generative Adversarial Network (Loop-GAN) is introduced to synthesize pseudo data to enhance the quality of synthetic data. Next, a bidirectional GRU model is structurally integrated with the Wiener process. In this design, the GRU model is configured in an Auto-Encoder-like structure, with the Wiener process serving as the hidden layer within this Auto-Encoder. Importantly, both the GRU and Wiener processes are jointly optimized with the assistance of Loop-GAN, emphasizing the collaborative nature of information fusion in this approach. The effectiveness of the proposed GRU-AE-Wiener is validated using the PHM 2012 dataset and XJTU-SY dataset. Experimental results underscore its superior RUL predictive performance compared to other deep learning models, highlighting the practical application of information fusion principles in bearing health assessment.

A novel temporal-frequency combination pattern optimization approach based on information fusion for motor imagery BCIs

Motor imagery (MI) stands as a powerful paradigm within Brain-Computer Interface (BCI) research due to its ability to induce changes in brain rhythms detectable through common spatial patterns (CSP). However, the raw feature sets captured often contain redundant and invalid information, potentially hindering CSP performance. Methodology-wise, we propose the Information Fusion for Optimizing Temporal-Frequency Combination Pattern (IFTFCP) algorithm to enhance raw feature optimization. Initially, preprocessed data undergoes simultaneous processing in both time and frequency domains via sliding overlapping time windows and filter banks. Subsequently, we introduce the Pearson-Fisher combinational method along with Discriminant Correlation Analysis (DCA) for joint feature selection and fusion. These steps aim to refine raw electroencephalogram (EEG) features. For precise classification of binary MI problems, an Radial Basis Function (RBF)-kernel Support Vector Machine classifier is trained. To validate the efficacy of IFTFCP and evaluate it against other techniques, we conducted experimental investigations using two EEG datasets. Results indicate a notably superior classification performance, boasting an average accuracy of 78.14% and 85.98% on dataset 1 and dataset 2, which is better than other methods outlined in this article. The study’s findings suggest potential benefits for the advancement of MI-based BCI strategies, particularly in the domain of feature fusion.

Multi-stream domain adversarial prototype network for integrated smart roller TBM main bearing fault diagnosis across various low rotating speeds

The well-being of tunnel boring machines (TBM) heavily relies on the health status of main bearings, significantly impacting both safety and operational efficiency. Investigating the operation monitoring and diagnosis algorithms of TBM main bearings in practical scenarios with limited fault data and real-time variations in low-speed heavy-load operating conditions poses a research challenge. To solve these issues, a novel domain-adversarial prototype network diagnostic algorithm with a multi-stream fusion feature encoder (MDAPN) based on smart roller monitoring data is proposed. In contrast to existing external vibration monitoring schemes for bearings, monitoring data from smart rollers can mitigate complex external interference and greatly shorten the transmission path of fault sources. The algorithm encompasses a multi-stream fusion feature encoder, domain discriminator, and prototype network classifier. Specifically, the multi-stream fusion feature encoder adopts the two-stream convolutional network to deeply extract the axial-radial fused features of rollers, incorporating shallow statistical feature information. Domain-invariant features are generated based on domain adversarial learning strategies, and the prototype network classifier reduces dependence on target domain samples. An integrated smart roller main bearing fault simulation testbed was built, conducting 6 sets of cross-domain experiments. The proposed method reached an average accuracy of 98.41 % under 5-shot, surpassing the baseline method by 6.57 % at least. This validates that the MDAPN based on smart roller state exhibits excellent fault recognition performance for the main bearing under varying operational conditions with scarce available samples.

Comprehensive analysis of the editing window of C-to-T TALE base editors

One of the most recent advances in the genome editing field has been the addition of “TALE Base Editors”, an innovative platform for cell therapy that relies on the deamination of cytidines within double strand DNA, leading to the formation of an uracil (U) intermediate. These molecular tools are fusions of transcription activator-like effector domains (TALE) for specific DNA sequence binding, split-DddA deaminase halves that will, upon catalytic domain reconstitution, initiate the conversion of a cytosine (C) to a thymine (T), and an uracil glycosylase inhibitor (UGI). We developed a high throughput screening strategy capable to probe key editing parameters in a precisely defined genomic context in cellulo, excluding or minimizing biases arising from different microenvironmental and/or epigenetic contexts. Here we aimed to further explore how target composition and TALEB architecture will impact the editing outcomes. We demonstrated how the nature of the linker between TALE array and split DddAtox head allows us to fine tune the editing window, also controlling possible bystander activity. Furthermore, we showed that both the TALEB architecture and spacer length separating the two TALE DNA binding regions impact the target TC editing dependence by the surrounding bases, leading to more restrictive or permissive editing profiles.

Double domain fusion improves the reverse transcriptase activity and inhibitor tolerance of Bst DNA polymerase

To enhance the DNA/RNA amplification efficiency and inhibitor tolerance of Bst DNA polymerase, four chimeric Bst DNA polymerase by fusing with a DNA-binding protein Sto7d and/or a highly hydrophobic protein Hp47 to Bst DNA polymerase large fragment. One of chimeric protein HpStBL exhibited highest inhibitor tolerance, which retained high active under 0.1 U/μL sodium heparin, 0.8 ng/μL humic acid, 2.5× SYBR Green I, 8 % (v/v) whole blood, 20 % (v/v) tissue, and 2.5 % (v/v) stool. Meanwhile, HpStBL showed highest sensitivity (93.75 %) to crude whole blood infected with the African swine fever virus. Moreover, HpStBL showed excellent reverse transcriptase activity in reverse transcription loop-mediated isothermal amplification, which could successfully detect 0.5 pg/μL severe acute respiratory syndrome coronavirus 2 RNA in the presence of 1 % (v/v) stools. The fusion of two domains with different functions to Bst DNA polymerase would be an effective strategy to improve Bst DNA polymerase performance in direct loop-mediated isothermal amplification and reverse transcription loop-mediated isothermal amplification detection, and HpStBL would be a promising DNA polymerase for direct African swine fever virus/severe acute respiratory syndrome coronavirus 2 detection due to simultaneously increased inhibitor tolerance and reverse transcriptase activity.

A novel subunit vaccine based on Fiber1/2 knob domain provides full protection against fowl adenovirus serotype 4 and induces stronger immune responses than a Fiber2 subunit vaccine

Outbreaks of hepatitis-hydropericardium syndrome (HHS) caused by fowl adenovirus serotype 4 (FAdV-4) have resulted in huge economic losses to the poultry industry in China since 2015. However, commercially available vaccines against the FAdV-4 infection remain scarce. In our study, subunit vaccine candidates derived from the bacterially expressed recombinant Fiber1 knob domain and Fiber2 knob domain fusion protein (termed as Fiber1/2 knob subunit vaccine) and Fiber2 protein (termed as Fiber2 subunit vaccine) of the FAdV-4 SDSX strain were developed. Immunogenicity evaluation showed that the Fiber1/2 knob subunit vaccine induced the production of antibodies at 7 d postvaccination (dpv), earlier than the Fiber2 subunit vaccine. Moreover, the neutralizing antibody level of the Fiber1/2 subunit vaccine group was higher than the Fiber2 subunit vaccine group, showing significant differences at 14, 21, and 28 dpv. Immune protection test results revealed that both Fiber1/2 knob subunit and Fiber2 subunit vaccines could protect chickens from death against FAdV-4 challenge, although the weight of chickens in the Fiber1/2 knob subunit vaccine group decreased less. Furthermore, analysis of plasma Glutamic oxaloacetic transaminase (AST) and blood glutamic pyruvic transaminase (ALT) levels suggested that the Fiber1/2 subunit vaccine can significantly inhibit liver damage caused by FAdV-4 infection and is more effective in blocking the pathogenicity of FAdV-4 in target organs. In addition, the Fiber1/2 knob subunit vaccine further reduced the viral load in different tissues and virus shedding in chickens than the Fiber2 subunit vaccine. Overall, the Fiber1/2 knob subunit vaccine was more effective than the Fiber2 subunit vaccine. These findings lay the foundation for the development of more effective FAdV-4 subunit vaccines.

Preserved structure and function of human serum albumin self-folded in the oxidative cytoplasm of Escherichia coli

Human serum albumin (HSA), a polypeptide featuring 17 disulfide bonds, acts as a crucial transport protein in human blood plasma. Its extended circulation half-life, mediated by FcRn (neonatal Fc receptor)-facilitated recycling, positions HSA as an excellent carrier for long-acting drug delivery. However, the conventional method of obtaining HSA from human blood faces limitations due to availability and potential contamination risks, such as blood-borne diseases. This study introduced SHuffle, an oxidative Escherichia coli (E. coli) expression system, for the production of recombinant HSA (rHSA) that spontaneously self-folds into its native conformation. This system ensures precise disulfide bond formation and correct folding of cysteine-rich rHSA, eliminating the need for chaperone co-expression or domain fusion of a folding enhancer. The purified rHSA underwent thorough physicochemical characterization, including mass spectrometry, circular dichroism spectroscopy, intrinsic fluorescence spectroscopy, esterase-like activity assay, and size exclusion chromatography, to assess critical quality attributes. Importantly, rHSA maintained native binding affinity to FcRn and the albumin-binding domain. Collectively, our analyses demonstrated a high comparability between rHSA and plasma-derived HSA. The expression of rHSA in E. coli with an oxidizing cytosol provides a secure and cost-effective approach, enhancing the potential of rHSA for diverse medical applications.

Cell coupling compensates for changes in single-cell Her6 dynamics and provides phenotypic robustness.

This paper investigates the effect of altering the protein expression dynamics of the bHLH transcription factor Her6 at the single-cell level in the embryonic zebrafish telencephalon. Using a homozygote endogenous Her6:Venus reporter and 4D single-cell tracking, we show that Her6 oscillates in neural telencephalic progenitors and that the fusion of protein destabilisation (PEST) domain alters its expression dynamics, causing most cells to downregulate Her6 prematurely. However, counterintuitively, oscillatory cells increase, with some expressing Her6 at high levels, resulting in increased heterogeneity of Her6 expression in the population. These tissue-level changes appear to be an emergent property of coupling between single-cells, as revealed by experimentally disrupting Notch signalling and by computationally modelling alterations in Her6 protein stability. Despite the profound differences in the single-cell Her6 dynamics, the size of the telencephalon is only transiently altered and differentiation markers do not exhibit significant differences early on; however, a small increase is observed at later developmental stages. Our study suggests that cell coupling provides a compensation strategy, whereby an almost normal phenotype is maintained even though single-cell gene expression dynamics are abnormal, granting phenotypic robustness.

Unsupervised multi-level spatio-spectral fusion transformer for hyperspectral image super-resolution

Fusing a low spatial resolution hyperspectral image (LR-HSI) with a high spatial resolution multispectral image (HR-MSI) is widely used for HSI super-resolution. Recent works still face problems in exploring global spatio-spectral correlation and lack effective utilization of multi-level features of the inputs (i.e., HR-MSI and LR-HSI), which results in a lack of similarity between the reconstruction and the inputs, ultimately causing significant spatio-spectral distortion. To solve these problems, we design an Unsupervised Multi-level Spatio-spectral Fusion Transformer (UMSFT). In UMSFT, a novel multi-level features fusion strategy is constructed, which fuses the hierarchy features of the inputs via proposed Spatio-spectral Feature Fusion Attention Blocks (S2F-FAB) in a level-by-level manner, thereby fully exploring the interaction between the hierarchical features. The S2F-FAB is specially designed for HSI-MSI fusion consisting of two components: (1) a spatial fusion module (Spa-FM) is designed for spatial domain fusion, and its output is set as Values (V) of a transformer; (2) a novel spectral feature cross attention (Spe- FCA) formulates the features of LR-HSI and HR-MSI as Queries (Q) and Keys (K), respectively, and achieves spectral domain fusion by applying attention mechanism along the spectral dimension. Incorporating spatial detail reconstruction and spectral feature integration into the attention mechanism, the S2F-FAB efficiently exploits the spatio-spectral correlation between target HR-HSI and inputs. Experimental results on three public datasets and our real-world images show the superiority of our method as compared with eleven state-of-the-art methods. Codes will be available at https://github.com/Caoxuheng/HIFtool.

FBN-TCN: Temporal convolutional neural network based on spatial domain fusion brain networks for affective brain–computer interfaces

Background and objective: Electroencephalogram (EEG) emotion recognition is increasingly popular for visualizing and quantifying brain networks, with potential applications in medical cognition and human–computer interaction. However, single brain space is limited, and multi-channel recognition poses challenges with calculation volume and information redundancy. Therefore, this paper focuses on studying the affective brain–computer interface (aBCI) and critical lead acquisition of the fusion brain network. Method: This paper employs two correlation algorithms to build functional and causal networks. A fusion brain network (FBN) was proposed to capture emotional information in the spatial domain. Additionally, a deep network recognition model (FBN-TCN) is introduced to capture the dynamic changes of spatial domain emotional information over time. A critical channel combination with high emotional correlation among multiple channels is obtained by the brain causal networks. All proposed models were performed on the SEED dataset through subject-dependent and subject-independent experiments. Results: Experimental results demonstrate that FBN can cover more emotional information to improve emotional recognition. The critical channel aBCI achieves comparable recognition performance to the full-channel aBCI, with recognition results of 95.51%/3.15% and 96.34%/1.62% in the Total band, respectively. The generalization and effectiveness of the model were verified through ablation experiments, which involved subject-dependent and subject-independent emotion recognition of DEAP, and emotion recognition of SEED on model block transformation. Conclusion and significance: This paper can simplify the electrode cap device and maximizes the utilization of EEG spatial domain emotional information. It is highly significant for enhancing the performance of aBCI in human–computer interaction.

Characterization of winged helix domain fusion endonucleases as N6-methyladenine-dependent type IV restriction systems.

Winged helix (wH) domains, also termed winged helix-turn-helix (wHTH) domains, are widespread in all kingdoms of life and have diverse roles. In the context of DNA binding and DNA modification sensing, some eukaryotic wH domains are known as sensors of non-methylated CpG. In contrast, the prokaryotic wH domains in DpnI and HhiV4I act as sensors of adenine methylation in the 6mApT (N6-methyladenine, 6mA, or N6mA) context. DNA-binding modes and interactions with the probed dinucleotide are vastly different in the two cases. Here, we show that the role of the wH domain as a sensor of adenine methylation is widespread in prokaryotes. We present previously uncharacterized examples of PD-(D/E)XK-wH (FcyTI, Psp4BI), PUA-wH-HNH (HtuIII), wH-GIY-YIG (Ahi29725I, Apa233I), and PLD-wH (Aba4572I, CbaI) fusion endonucleases that sense adenine methylation in the Dam+ Gm6ATC sequence contexts. Representatives of the wH domain endonuclease fusion families with the exception of the PLD-wH family could be purified, and an in vitro preference for adenine methylation in the Dam context could be demonstrated. Like most other modification-dependent restriction endonucleases (MDREs, also called type IV restriction systems), the new fusion endonucleases except those in the PD-(D/E)XK-wH family cleave close to but outside the recognition sequence. Taken together, our data illustrate the widespread combinatorial use of prokaryotic wH domains as adenine methylation readers. Other potential 6mA sensors in modified DNA are also discussed.

Genome mining of labdane-related diterpenoids: Discovery of the two-enzyme pathway leading to (-)-sandaracopimaradiene in the fungus Arthrinium sacchari.

Labdane-related diterpenoids (LRDs) in fungi are a pharmaceutically important, but underexplored family of natural products. In the biosynthesis of fungal LRDs, bifunctional terpene cyclases (TCs) consisting of αβγ domains are generally used to synthesize the polycyclic skeletones of LRDs. Herein, we conducted genome mining of LRDs in our fungal genome database and identified a unique pair of TCs, AsPS and AsCPS, in the fungus Arthrinium sacchari. AsPS consists of catalytically active α and inactive β domains, whereas AsCPS contains βγ domains and a truncated α domain. Heterologous expression in Aspergillus oryzae and biochemical characterization of recombinant proteins demonstrated that AsCPS synthesized copalyl diphosphate and that AsPS then converted it to (-)-sandaracopimaradiene. Since AsPS and AsCPS have distinct domain organizations from those of known fungal TCs and are likely generated through fusion or loss of catalytic domains, our findings provide insight into the evolution of TCs in fungi.

N-terminal domain truncation yielded a unique dimer of polysaccharide hydrolase with enhanced enzymatic activity, stability and calcium ion independence

Domain engineering, including domain truncation, fusion, or swapping, has become a common strategy to improve properties of enzymes, especially glycosyl hydrolases. However, there are few reports explaining the mechanism of increased activity from a protein structure perspective. Amy703 is an alkaline amylase with a unique N-terminal domain. Prior studies have shown that N-Amy, a mutant without an N-terminal domain, exhibits improved activity, stability, and calcium ion independence. In this study, we have used X-ray crystallography to determine the crystal structure of N-Amy and used AlphaFold2 to model the Amy703 structure, respectively. We further used size exclusion chromatography to show that Amy703 existed as a monomer, whereas N-Amy formed a unique dimer. It was found that the N-terminus of one monomer of N-Amy was inserted into the catalytic domain of its symmetrical subunit, resulting in the expansion of the catalytic pocket. This also significantly increased the pKa of the hydrogen donor Glu350, thereby enhancing substrate binding affinity and contributing to increased N-Amy activity. Meanwhile, two calcium ions were found to bind to N-Amy at different binding sites, which also contributed to the stability of protein. Therefore, this study provided new structural insights into the mechanisms of various glycosyl hydrolases.

Modular design of bi- and multi-specific knob domain fusions.

The therapeutic potential of bispecific antibodies is becoming widely recognised, with over a hundred formats already described. For many applications, enhanced tissue penetration is sought, so bispecifics with low molecular weight may offer a route to enhanced potency. Here we report the design of bi- and tri-specific antibody-based constructs with molecular weights as low as 14.5 and 22 kDa respectively. Autonomous bovine ultra-long CDR H3 (knob domain peptide) modules have been engineered with artificial coiled-coil stalks derived from Sin Nombre orthohantavirus nucleocapsid protein and human Beclin-1, and joined in series to produce bi- and tri-specific antibody-based constructs with exceptionally low molecular weights. Knob domain peptides with coiled-coil stalks retain high, independent antigen binding affinity, exhibit exceptional levels of thermal stability, and can be readily joined head-to-tail yielding the smallest described multi-specific antibody format. The resulting constructs are able to bind simultaneously to all their targets with no interference. Compared to existing bispecific formats, the reduced molecular weight of the knob domain fusions may enable enhanced tissue penetration and facilitate binding to cryptic epitopes that are inaccessible to conventional antibodies. Furthermore, they can be easily produced at high yield as recombinant products and are free from the heavy-light chain mispairing issue. Taken together, our approach offers an efficient route to modular construction of minimalistic bi- and multi-specifics, thereby further broadening the therapeutic scope for knob domain peptides.

FoSp: Focus and Separation Network for Early Smoke Segmentation

Early smoke segmentation (ESS) enables the accurate identification of smoke sources, facilitating the prompt extinguishing of fires and preventing large-scale gas leaks. But ESS poses greater challenges than conventional object and regular smoke segmentation due to its small scale and transparent appearance, which can result in high miss detection rate and low precision. To address these issues, a Focus and Separation Network (FoSp) is proposed. We first introduce a Focus module employing bidirectional cascade which guides low-resolution and high-resolution features towards mid-resolution to locate and determine the scope of smoke, reducing the miss detection rate. Next, we propose a Separation module that separates smoke images into a pure smoke foreground and a smoke-free background, enhancing the contrast between smoke and background fundamentally, improving segmentation precision. Finally, a Domain Fusion module is developed to integrate the distinctive features of the two modules which can balance recall and precision to achieve high F_beta. Futhermore, to promote the development of ESS, we introduce a high-quality real-world dataset called SmokeSeg, which contains more small and transparent smoke than the existing datasets. Experimental results show that our model achieves the best performance on three available smoke segmentation datasets: SYN70K (mIoU: 83.00%), SMOKE5K (F_beta: 81.6%) and SmokeSeg (F_beta: 72.05%). The code can be found at https://github.com/LujianYao/FoSp.