Sequence Reconstruction Research Articles

Deep learning reconstruction of zero-echo time sequences to improve visualization of osseous structures and associated pathologies in MRI of cervical spine

ObjectivesTo determine whether deep learning-based reconstructions of zero-echo-time (ZTE-DL) sequences enhance image quality and bone visualization in cervical spine MRI compared to traditional zero-echo-time (ZTE) techniques, and to assess the added value of ZTE-DL sequences alongside standard cervical spine MRI for comprehensive pathology evaluation.MethodsIn this retrospective study, 52 patients underwent cervical spine MRI using ZTE, ZTE-DL, and T2-weighted 3D sequences on a 1.5-Tesla scanner. ZTE-DL sequences were reconstructed from raw data using the AirReconDL algorithm. Three blinded readers independently evaluated image quality, artifacts, and bone delineation on a 5-point Likert scale. Cervical structures and pathologies, including soft tissue and bone components in spinal canal and neural foraminal stenosis, were analyzed. Image quality was quantitatively assessed by signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR).ResultsMean image quality scores were 2.0 ± 0.7 for ZTE and 3.2 ± 0.6 for ZTE-DL, with ZTE-DL exhibiting fewer artifacts and superior bone delineation. Significant differences were observed between T2-weighted and ZTE-DL sequences for evaluating intervertebral space, anterior osteophytes, spinal canal, and neural foraminal stenosis (p < 0.05), with ZTE-DL providing more accurate assessments. ZTE-DL also showed improved evaluation of the osseous components of neural foraminal stenosis compared to ZTE (p < 0.05).ConclusionsZTE-DL sequences offer superior image quality and bone visualization compared to ZTE sequences and enhance standard cervical spine MRI in assessing bone involvement in spinal canal and neural foraminal stenosis.Critical relevance statementDeep learning-based reconstructions improve zero-echo-time sequences in cervical spine MRI by enhancing image quality and bone visualization. This advancement offers additional insights for assessing bone involvement in spinal canal and neural foraminal stenosis, advancing clinical radiology practice.Key PointsConventional MRI encounters challenges with osseous structures due to low signal-to-noise ratio.Zero-echo-time (ZET) sequences offer CT-like images of the C-spine but with lower quality.Deep learning reconstructions improve image quality of zero-echo-time sequences.ZTE sequences with deep learning reconstructions refine cervical spine osseous pathology assessment.These sequences aid assessment of bone involvement in spinal and foraminal stenosis.Graphical

Ancestral Sequence Reconstruction and Comprehensive Computational Simulations Unmask an Efficient PET Hydrolase with the Wobbled Catalytic Triad.

Beyond directed evolution, ancestral sequence reconstruction (ASR) has emerged as a powerful strategy for engineering proteins with superior functional properties. Herein, we harnessed ASR to uncover robust PET hydrolase variants, expanding the repertoire of PET-degrading enzymes and providing deeper insights into the underlying mechanisms of PET hydrolysis. As a result, ASR1-PETase, featuring a unique cysteine catalytic site, was discovered. Despite having only 19.3% sequence identity with IsPETase, ASR1-PETase demonstrated improved PET degradation efficiency, with a finely-tuned substrate-binding cleft. Comprehensive experimental validation, including mutagenesis studies and comparisons with six state-of-the-art PET hydrolases, combined with microsecond-scale molecular dynamics (MD) simulations and QM-cluster calculations, revealed that ASR1-PETase's C161 catalytic residue assisted with the wobbled H242 can simultaneously cleave both ester bonds of BHET-a feature not commonly observed in other PET hydrolases. This mechanism may serve as the primary driving force for accelerating PET hydrolysis while minimizing the accumulation of the intermediate MHET, thereby enhancing the efficiency of TPA production.

Ancestral sequence reconstruction of the Mic60 Mitofilin domain reveals residues supporting respiration in yeast.

In eukaryotes, cellular respiration takes place in the cristae of mitochondria. The mitochondrial inner membrane protein Mic60, a core component of the mitochondrial contact site and cristae organizing system (MICOS), is crucial for the organization and stabilization of crista junctions and its associated functions. While the C-terminal Mitofilin domain of Mic60 is necessary for cellular respiration, the sequence determinants for this function have remained unclear. Here, we used ancestral sequence reconstruction to generate Mitofilin ancestors up to and including the last opisthokont common ancestor (LOCA). We found that yeast-lineage derived Mitofilin ancestors as far back as the LOCA rescue respiration. By comparing Mitofilin ancestors, we identified four residues sufficient to explain the respiratory difference between yeast- and animal-derived Mitofilin ancestors. Our results provide a foundation for investigating the conservation of Mic60-mediated cristae junction interactions.

Ancestral carbonic anhydrase with significantly enhanced stability and activity for CO2 capture and utilization.

Carbonic anhydrases (CAs) has garnered increasing attention in carbon capture, utilization and storage (CCUS) due to their ecological friendliness. However, most of them suffer susceptibility to deactivation in harsh conditions. Herein, a reliable dataset was adopted for creating ancestral CAs through an optimized ancestral sequence reconstruction (ASR) method. After prescreening, the ancestor AncCA19 was obtained and successfully expressed. The hydration activity of AncCA19 was as high as 58,859 WAU/mg, with the optimum temperature and pH obtained by esterase assay at 100 ℃ and 9, respectively. AncCA19 had the longest half-life (1.7 h) at 95 ℃ compared with existing CAs. After 2 weeks' incubation in artificial seawater at 30 ℃ or 25.0 % N-methyldiethanolamine (MDEA) at 60 ℃, the activities remained above 47,370 WAU/mg and 6,596 WAU/mg, respectively. Thus, AncCA19, as a novel benchmark of CAs, exhibits exceptional stability in a variety CCUS applications, establishing a versatile candidate for effective CO2 capture.

SRdetector: Sequence Reconstruction Method for Microservice Anomaly Detection

With the expansion of microservice-based applications over time, the number of microservices rises, resulting in an augmentation of the volume of performance metrics. Consequently, selecting the appropriate performance metrics for anomaly detection becomes a critical challenge. Since these performance metrics are typically strongly correlated with timestamps, they form time series data comprising timestamp–value pairs. To address this, we propose SRdetector, a feature-enhanced Transformer-based model that adopts a time series forecasting approach to detect anomalies in microservices. Furthermore, we integrate a dynamic weight adjustment mechanism into the original Transformer attention mechanism to assign weights to different performance and temporal features. This enables the model to dynamically learn the significance of various features at different time intervals, effectively serving as a feature selection method for microservice performance metrics. Finally, anomaly detection in microservices is conducted by evaluating the predicted performance metric data based on confidence intervals.

The Presence of Two Distinct Lineages of the Foot-And-Mouth Disease Virus Type A in Russia in 2013-2014 Has Significant Implications for the Epidemiology of the Virus in the Region.

Molecular surveillance of FMD epidemiology is a fundamental tool for advancing our understanding of virus biology, monitoring virus evolution, and guiding vaccine design. The accessibility of genetic data will facilitate a more comprehensive delineation of FMDV phylogeny on a global scale. In this study, we investigated the FMDV strains circulating in Russia during the 2013-2014 period in geographically distant regions utilizing whole genome sequencing followed by maximum-likelihood phylogenetic reconstruction of whole genome and VP1 gene sequences. Phylogenetic analysis showed congruence in the topology of the phylogenetic trees constructed using the complete genome and VP1 gene sequence, clearly demonstrating that the isolates analyzed belong to two distinct genetic lineages: A/SEA97 in the Far East and Iran-05 in the North Caucasus. The A/SEA97 isolates exhibited a close genetic identity to those from China and Mongolia, whereas the Iran-05 isolates demonstrated clusterization with those from Turkey. The vaccine-matching studies with isolates from the Far East and North Caucasus revealed no antigenic homology with A/SEA-97 (r1 = 0.015-0.29) and A/Iran 05 (r1 = 0.009-0.17). The close genetic relationship of FMDV in the reported outbreak waves to those from neighboring countries indicates that animal movement could contribute to spillover and virus dispersal. The phylogenetic data reported here provide insight into the molecular epidemiology of FMD in the Eurasia region, elucidating the circulation pattern, molecular evolution, and genetic diversity, which is highly valuable for guiding vaccine designs and improving regional eradication policies.

Ancestral Sequence Reconstruction Reveals Determinants of Regioselectivity in C(sp3)-H Oxyfunctionalization Reactions by CYP505Es

Ancestral Sequence Reconstruction Reveals Determinants of Regioselectivity in C(sp³)-H Oxyfunctionalization Reactions by CYP505Es

Robustness of ancestral sequence reconstruction to among-site evolutionary heterogeneity and epistasis.

Ancestral sequence reconstruction (ASR) is typically performed using homogeneous evolutionary models, which assume that the same substitution propensities affect all sites and lineages. These assumptions are routinely violated: heterogeneous structural and functional constraints favor different amino acid states at different sites, and these constraints often change among lineages as epistatic substitutions accrue at other sites. To evaluate how realistic violations of the homogeneity assumption affect ASR, we developed site-specific substitution models and parameterized them using data from deep mutational scanning experiments on three protein families; we then used these models to perform ASR on the empirical alignments and on alignments simulated under heterogeneous conditions derived from the experiments. Extensive among-site and -lineage heterogeneity is present in these datasets, but the sequences reconstructed from empirical alignments are almost identical, irrespective of whether heterogeneous or homogeneous models are used for ASR. The rare differences occur primarily when phylogenetic signal is weak - at fast-evolving sites and nodes connected by long branches. When ASR is performed on simulated data, errors in the reconstructed sequences become more likely as branch lengths increase, but incorporating heterogeneity into the model does not improve accuracy. These data establish that ASR is robust to unincorporated realistic forms of evolutionary heterogeneity, because the primary determinant of ASR is phylogenetic signal, not the substitution model. The best way to improve accuracy is therefore not to develop more elaborate models but to apply ASR to densely sampled alignments that maximize phylogenetic signal at the nodes of interest.

Leveraging ancestral sequence reconstruction for protein representation learning

Leveraging ancestral sequence reconstruction for protein representation learning

Design of ancestral mammalian alkaline phosphatase bearing high stability and productivity.

Sequence-based protein redesign methods, such as full consensus design (FCD) and ancestral sequence reconstruction (ASR), have the potential to construct new enzymes utilizing protein sequence data registered in a rapidly expanding sequence database. The high thermostability of these enzymes would expand their application in diagnostics and molecular biology. These enzymes have also demonstrated a high level of active expression by Escherichia coli. These characteristics make these APs attractive candidates for industrial application. In addition, different amino acid residues are primarily responsible for thermal stability and active expression, suggesting important implications for strategies for designing enzymes by FCD and ASR.

Stable and Promiscuous Galactose Oxidases Engineered by Directed Evolution, Atomistic Design, and Ancestral Sequence Reconstruction.

Galactose oxidase (GOase) is a versatile biocatalyst with a wide range of potential applications, ranging from synthetic chemistry to bioelectrochemical devices. Previous GOase engineering by directed evolution generated the M-RQW mutant, with unprecedented new-to-nature oxidation activity at the C6-OH group of glucose, and a mutational backbone that helped to unlock its promiscuity toward other molecules, including secondary alcohols. In the current study, we have used the M-RQW mutant as a starting point to engineer a set of GOases that are very thermostable and that are easily produced at high titers in yeast, enzymes with latent activities applicable to sustainable chemistry. To boost the generation of sequence and functional diversity, the directed evolution workflow incorporated one-shot computational mutagenesis by the PROSS algorithm and ancestral sequence reconstruction. This synergetic approach helped produce a rapid rise in functional expression by Pichia pastoris, achieving g/L production in a fed-batch bioreactor while the different GOases designed were resistant to pH and high temperature, with T50 enhancements up to 27 °C over the parental M-RQW. These designs displayed latent activity against glucose and an array of secondary aromatic alcohols with different degrees of bulkiness, becoming a suitable point of departure for the future engineering of industrial GOases.

Engineering the Reaction Pathway of a Non-heme Iron Oxygenase Using Ancestral Sequence Reconstruction.

Non-heme iron (FeII), α-ketoglutarate (α-KG)-dependent oxygenases are a family of enzymes that catalyze an array of transformations that cascade forward after the formation of radical intermediates. Achieving control over the reaction pathway is highly valuable and a necessary step toward broadening the applications of these biocatalysts. Numerous approaches have been used to engineer the reaction pathway of FeII/α-KG-dependent enzymes, including site-directed mutagenesis, DNA shuffling, and site-saturation mutagenesis, among others. Herein, we showcase a novel ancestral sequence reconstruction (ASR)-guided strategy in which evolutionary information is used to pinpoint the residues critical for controlling different reaction pathways. Following this, a combinatorial site-directed mutagenesis approach was used to quickly evaluate the importance of each residue. These results were validated using a DNA shuffling strategy and through quantum mechanical/molecular mechanical (QM/MM) simulations. Using this approach, we identified a set of active site residues together with a key hydrogen bond between the substrate and an active site residue, which are crucial for dictating the dominant reaction pathway. Ultimately, we successfully converted both extant and ancestral enzymes that perform benzylic hydroxylation into variants that can catalyze an oxidative ring-expansion reaction, showcasing the potential of utilizing ASR to accelerate the reaction pathway engineering within enzyme families that share common structural and mechanistic features.

Protein Structural Modeling and Transport Thermodynamics Reveal That Plant Cation-Chloride Cotransporters Mediate Potassium-Chloride Symport.

Plant cation-chloride cotransporters (CCCs) are proposed to be Na+-K+-2Cl- transporting membrane proteins, although evolutionarily, they associate more closely with K+-Cl- cotransporters (KCCs). Here, we investigated grapevine (Vitis vinifera L.) VvCCC using 3D protein modeling, bioinformatics, and electrophysiology with a heterologously expressed protein. The 3D protein modeling revealed that the signatures of ion binding sites in plant CCCs resembled those of animal KCCs, which was supported by phylogenomic analyses and ancestral sequence reconstruction. The conserved features of plant CCCs and animal KCCs included predicted K+ and Cl--binding sites and the absence of a Na+-binding site. Measurements with VvCCC-injected Xenopus laevis oocytes with VvCCC localizing to plasma membranes indicated that the oocytes had depleted intracellular Cl- and net 86Rb fluxes, which agreed with thermodynamic predictions for KCC cotransport. The 86Rb uptake by VvCCC-injected oocytes was Cl--dependent, did not require external Na+, and was partially inhibited by the non-specific CCC-blocker bumetanide, implying that these properties are typical of KCC transporters. A loop diuretic-insensitive Na+ conductance in VvCCC-injected oocytes may account for earlier observations of Na+ uptake by plant CCC proteins expressed in oocytes. Our data suggest plant CCC membrane proteins are likely to function as K+-Cl- cotransporters, which opens the avenues to define their biophysical properties and roles in plant physiology.

Sequence signal prediction and reconstruction for multi-energy load forecasting in integrated energy systems: A bi-level multi-task learning method

Multi-energy load forecasting is the basis for the optimization and scheduling of integrated energy systems (IES). An IES contains much heterogeneous energy with volatility and temporality, and complex coupling relationships between the energy sources, which makes it difficult to accurately predict multiple loads simultaneously. Considering these issues, this paper proposes a bi-level multi-task learning (BiMTL) method for simultaneous sequence signal prediction and reconstruction via deep neural networks. In the feature construction process, considering the volatility and randomness of the load, the improved wavelet packet decomposition model is constructed to decompose the multiple load sequences into sub-signal with more obvious signal characteristics, and improves the ability of the forecasting model to capture and respond to the fluctuations of load sequences within a local range. In the modeling process, considering the excessive dependence of traditional signal reconstruction methods on the prediction results of sub-bands, this paper innovatively proposes a BiMTL method. This method employs a second-layer shared learning to perform signal reconstruction and error correction on the output of the first layer’s sub-band predictions. Finally, a BiMTL-LSTM model is constructed to extract the coupling and temporality of loads based on the LSTM hard-sharing mechanism. Experimental results show that the proposed method outperforms the current state-of-the-art methods on the widely used Tempe dataset, achieving an accuracy of up to 98.70%.

Evolutionary insights into the stereoselectivity of imine reductases based on ancestral sequence reconstruction

The stereoselectivity of enzymes plays a central role in asymmetric biocatalytic reactions, but there remains a dearth of evolution-driven biochemistry studies investigating the evolutionary trajectory of this vital property. Imine reductases (IREDs) are one such enzyme that possesses excellent stereoselectivity, and stereocomplementary members are pervasive in the family. However, the regulatory mechanism behind stereocomplementarity remains cryptic. Herein, we reconstruct a panel of active ancestral IREDs and trace the evolution of stereoselectivity from ancestors to extant IREDs. Combined with coevolution analysis, we reveal six historical mutations capable of recapitulating stereoselectivity evolution. An investigation of the mechanism with X-ray crystallography shows that they collectively reshape the substrate-binding pocket to regulate stereoselectivity inversion. In addition, we construct an empirical fitness landscape and discover that epistasis is prevalent in stereoselectivity evolution. Our findings emphasize the power of ASR in circumventing the time-consuming large-scale mutagenesis library screening for identifying mutations that change functions and support a Darwinian premise from a molecular perspective that the evolution of biological functions is a stepwise process.

Evolutionary analysis of Quinone Reductases 1 and 2 suggests that NQO2 evolved to function as a pseudoenzyme.

Quinone reductases 1 and 2 (NQO1 and NQO2) are paralogous FAD-linked enzymes found in all amniotes. NQO1 and NQO2 have similar structures, and both catalyze the reduction of quinones and other electrophiles; however, the two enzymes differ in their cosubstrate preference. While NQO1 can use both redox couples NADH and NADPH, NQO2 is almost inactive with these cosubstrates and instead must use dihydronicotinamide riboside (NRH) and small synthetic cosubstrates such as N-benzyl-dihydronicotinamide (BNAH) for efficient catalysis. We used ancestral sequence reconstruction to investigate the catalytic properties of a predicted common ancestor and two additional ancestors from each of the evolutionary pathways to extant NQO1 and NQO2. In all cases, the small nicotinamide cosubstrates NRH and BNAH were good cosubstrates for the common ancestor and the enzymes along both the NQO1 and NQO2 lineages. In contrast, with NADH as cosubstrate, extant NQO1 evolved to a catalytic efficiency 100 times higher than the common ancestor, while NQO2 has evolved to a catalytic efficiency 3000 times lower than the common ancestor. The evolutionary analysis combined with site-directed mutagenesis revealed a potential site of interaction for the ADP portion of NAD(P)H in NQO1 that is altered in charge and structure in NQO2. The results indicate that while NQO1 evolved to have greater efficiency with NAD(P)H, befitting an enzymatic function in cells, NQO2 was under selective pressure to acquire extremely low catalytic efficiency with NAD(P)H. These divergent trajectories have implications for the functions of both enzymes.

Ancestral Sequence Reconstruction Meets Machine Learning: Ene Reductase Thermostabilization Yields Enzymes with Improved Reactivity Profiles

Ancestral Sequence Reconstruction Meets Machine Learning: Ene Reductase Thermostabilization Yields Enzymes with Improved Reactivity Profiles

Masked Graph Autoencoder for Self‐Supervised Transportation Mode Recognition

ABSTRACTTransportation mode recognition reveals the traveling patterns of urban residents and plays an important role in travel analysis and city planning. Research has been done to identify transportation modes using deep learning methods. However, large‐scale and far‐ranging labeled trajectory data are hard to acquire, which hinders the training and application of supervised models. In this study, we propose a self‐supervised method for trajectory transportation modes recognition. First, we construct sequence and dependency graph for trajectory points. Next, we apply mask and remask strategy to graph node features to enhance its learning ability in semantic information. Notably, in order to learn features without labeled data, we introduce graph autoencoder into the framework. This method is able to learn the sequence and dependency feature reconstruction of the travel paths. It can obtain underlying semantic information with the mask and remask strategy. We applied our model to GPS trajectory dataset from the Microsoft GeoLife Project, achieving an average accuracy of 68.26% in identifying the transportation modes among seven categories. The result indicates that the proposed method offers a solution to identifying transportation modes without abundant labeled data. This method can also help to extract space–time features at different scales from trajectory data, which could be applied to various downstream tasks.

MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection.

A large number of sensors are typically installed in industrial plants to collect real-time operational data. These sensors monitor data with time series correlation and spatial correlation over time. In previous studies, GNN has built many successful models to deal with time series data, but most of these models have fixed perspectives and struggle to capture the dynamic correlations in time and space simultaneously. Therefore, this paper constructs a multi-scale dynamic graph neural network (MSDG) for anomaly detection in industrial sensor data. First, a multi-scale sliding window mechanism is proposed to input different scale sensor data into the corresponding network. Then, a dynamic graph neural network is constructed to capture the spatial-temporal dependencies of multivariate sensor data. Finally, the model comprehensively considers the extracted features for sequence reconstruction and utilizes the reconstruction errors for anomaly detection. Experiments have been conducted on three real public datasets, and the results show that the proposed method outperforms the mainstream methods.

Ancestor-originated engineering of medium-chain alcohol dehydrogenase enhances its catalytic compatibility by balancing activity, stability, and selectivity trade-offs

The role of medium-chain alcohol dehydrogenases (MDR) in the preparation of chiral drug intermediates is indispensable; however, limited activity and stability towards unnatural substrates are crucial factors that restrict their application. Further, research on the ancestral MDRs remains unclear. Therefore, enhancing the activity, selectivity, and stability of MDRs is necessary. In this study, we resurrected ancestors of the MDR family using ancestral sequence reconstruction with descendant GcADH as a probe. Among these ancestors, ancestor 136 exhibited increased thermal stability (ΔTm = 19.5 °C) and had a comparable conversion and selectivity as GcADH to the selected substrates. After in-silico screening using the Funclib tool, we obtained the mutant anc136-R2, which exhibited strengthened selectivity and activity towards N-Benzyl-3,3-difluoropiperidin-4-one (1a). Moreover, the mutations of anc136-R2 were simultaneously mapped to GcADH, resulting GcADH-R2. Based on the results of substrate spectrum characterisation, anc136 was more receptive to mutations. Specifically, for the substrates N-Boc-3-piperidone (4a) and N-boc-3-oxoazepane (7a), the conversion rates of anc136-R2 were 29 and 57 times higher, respectively, than those of GcADH-R2, without compromising selectivity. In conclusion, this work provides guidance for exploring the evolutionary trajectory of the MDR family and further illustrates the superior balanced “trade-off” of activity–selectivity–stability on ancestral enzymes.