Search

The discovery of new antimalarial drugs can be developed using asynchronized Plasmodium berghei malaria parasites in vivo in mice. Studies on a particular stage are also required to assess the effectiveness and mode of action of drugs. In this report, we used endoperoxide 6-(1,2,6,7-tetraoxaspiro [7.11] nonadec-4-yl) hexan-1-ol (N-251) as a model antimalarial compound on P. chabaudi parasites. We examined the antimalarial effect of N-251 against ring-stage- and trophozoite-stage-rich P. chabaudi parasites and asynchronized P. berghei parasites using the 4-day suppressive test. The ED50 values were 27, 22, and 22 mg/kg, respectively, and the antimalarial activity of N-251 was verified in both rodent malaria parasites. To assess the stage-specific effect of N-251 in vivo, we evaluated the change of parasitemia and distribution of parasite stages using ring-stage- and trophozoite-stage-rich P. chabaudi parasites with one-day drug administration for one life cycle. We discovered that the parasitemias decreased after 13 and 9 hours post-treatment in the ring-stage- and trophozoite-stage-rich groups, respectively. Additionally, in the ring-stage-rich N-251 treated group, the ring-stage parasites hindered trophozoite parasite development. For the trophozoite-stage-rich N-251 treated group, the distribution of the trophozoite stage was maintained without a change in parasitemia until 9 hours. Because of these findings, it can be concluded that N-251 suppressed the trophozoite stage but not the ring stage. We report for the first time that N-251 specifically suppresses the trophozoite stage using P. chabaudi in mice. The results show that P. chabaudi is a reliable model for the characterization of stage-specific antimalarial effects.

Recent 2D-to-3D human pose estimation (HPE) utilizes temporal consistency across sequences to alleviate the depth ambiguity problem but ignore the action related prior knowledge hidden in the pose sequence. In this paper, we propose a plug-and-play module named Action Prompt Module (APM) that effectively mines different kinds of action clues for 3D HPE. The highlight is that, the mining scheme of APM can be widely adapted to different frameworks and bring consistent benefits. Specifically, we first present a novel Action-related Text Prompt module (ATP) that directly embeds action labels and transfers the rich language information in the label to the pose sequence. Besides, we further introduce Action-specific Pose Prompt module (APP) to mine the position-aware pose pattern of each action, and exploit the correlation between the mined patterns and input pose sequence for further pose refinement. Experiments show that APM can improve the performance of most video-based 2D-to-3D HPE frameworks by a large margin., 6 pages, 4 figures, 2023ICME

Picking up multiple objects at once is a grasping skill that makes a human worker efficient in many domains. This paper presents a system to pick a requested number of objects by only picking once (OPO). The proposed Only-Pick-Once System (OPOS) contains several graph-based algorithms that convert the layout of objects into a graph, cluster nodes in the graph, rank and select candidate clusters based on their topology. OPOS also has a multi-object picking predictor based on a convolutional neural network for estimating how many objects would be picked up with a given gripper location and orientation. This paper presents four evaluation metrics and three protocols to evaluate the proposed OPOS. The results show OPOS has very high success rates for two and three objects when only picking once. Using OPOS can significantly outperform two to three times single object picking in terms of efficiency. The results also show OPOS can generalize to unseen size and shape objects.

Purpose: To demonstrate the prognostic value of pretreatment red blood cell distribution width (RDW) in patients with urological cancer. Methods: We searched the relevant literature on Web of Science, Cochrane Central Register of Controlled Trials, PubMed, Embase, Sinomed databases and Chinese National Knowledge Infrastructure up to 30 March 2022, to investigate the relationship between RDW levels and the prognosis of patients with urological tumors. Results: This study comprised 15 retrospective studies involving 9492 patients. Increased pretreatment RDW was associated with poorer overall survival (hazard ratio [HR]: 1.52; 95% CI: 1.27–1.82; p

A 31-year-old woman with transthyretin (TTR) amyloidosis secondary to a Thr60Ala mutation developed recurrent stroke-like episodes with fluctuating mental status. Evaluation for stroke and seizures was unrevealing. She was found to have leptomeningeal contrast enhancement on magnetic resonance imaging, which was confirmed to be CNS TTR amyloidosis on histopathology following brain and dura biopsy. While leptomeningeal disease has rarely been known to be associated with TTR amyloidosis, this is the first documented case of leptomeningeal disease secondary to a Thr60Ala mutation in the TTR gene. A literature review of TTR amyloidosis is presented with special focus on the treatment of leptomeningeal TTR amyloidosis.

Vincristine (VCR) is a chemotherapeutic agent widely used in treatment of malignancies. However, VCR has a limitation in use since it commonly causes a painful neuropathy (VCR-induced peripheral neuropathy, VIPN). Inflammatory cytokines secreted by immune cells such as macrophages can exacerbate allodynia and hyperalgesia, because inhibiting the inflammatory response is a treatment target for VIPN. In this study, we investigated whether Trichinella spiralis, a widely studied helminth for its immunomodulatory abilities, can alleviate VCR-induced allodynia. Von Frey test showed that T. spiralis infection improved mechanical allodynia at 10 days after VCR injection. We further observed whether the difference was due to mitigated axon degeneration, but no significant difference between the groups in axonal degeneration in sciatic nerves and intra-epidermal nerve fibers was found. Conversely, we observed that number of infiltrated macrophages was decreased in the sciatic nerves of the T. spiralis infected mice. Moreover, treatment of T. spiralis excretory-secretory products caused peritoneal macrophages to secrete decreased level of IL-1β. This study suggests that T. spiralis can relieve VCR-induced mechanical allodynia by suppressing neuroinflammation and that application of controllable degree of helminth may prove beneficial for VIPN treatment.

The shiitake mushroom is the most commonly cultivated edible mushroom in the world, and is rich in protein. This study aims to obtain the peptides with α-glucosidase inhibition activity from shiitake mushroom protein hydrolysate. The conditions of enzymatic hydrolysis of shiitake mushroom protein were optimized by response surface test. The results showed that the optimal conditions were as follows: the E/S was 3390 U/g, the solid–liquid ratio was 1:20, the hydrolysis temperature and time were 46 °C and 3.4 h, respectively, and the pH was 7. The active peptides were separated by gel filtration and identified by LC-MS/MS analysis and virtual screening. The results indicated that fourteen peptides were identified by LC-MS/MS. Among them, four new peptides (EGEPKLP, KDDLRSP, TPELKL, and LDYGKL) with the higher docking score were selected and chemically synthesized to verify their inhibition activity. The IC50 values of EGEPKLP, KDDLRSP, TPELKL, and LDYGKL for α-glucosidase inhibition activity ranged from 452 ± 36 μmol/L to 696 ± 39 μmol/L. The molecular docking results showed that the hydrogen bond and arene–cation bond were the two major interactions between four peptides and 2QMJ. The hydrogen bonds were crucial to the inhibition activity of α-glucosidase. The results indicate the potential of using the peptides from shiitake mushroom protein as functional food with α-glucosidase inhibition activity.

Background: With the acceleration of urbanization, the demand for people to return to forests and their desire for nature is also increasing annually. However, whether the visual behaviors and aesthetic preferences of users vary with background attributes when viewing forest landscapes remains to be further explored. This information will help forest landscape planners and managers to create a forest landscape space suitable for different groups of people. Purpose: This study mainly discusses whether people’s professional background will affect their visual behavior, satisfaction preference and re-viewing intention of a landscape and discusses the relationship among them. Main results: (1) Under the background of an in-forest landscape, the visual behavior of users in different layouts presents great similarity. (2) Satisfaction preference for landscapes is not significantly influenced by the user’s background, but there is a significant linear relationship between satisfaction preference and re-viewing intention (Adj. R2 = 0.412 − 0.697, sig. = 0.000). (3) The spatial cognition that affects users’ visual behavior and satisfaction preference in a scene will change with the landscape layout and users’ professional background. Therefore, we suggest that landscape planners and managers should consider not only the spatial layout of the landscape itself but also the user’s own attributes (whether they have professional knowledge of the landscape) when optimizing the visual behavior and satisfaction preference of in-forest landscape space. This will ensure that users have a high sense of identity and attachment to the scene in a targeted way, thus arousing a greater “desire” to re-view the scene. In addition, our results can identify a more satisfactory course for tour routes according to the user’s professional background, thus enhancing the rate of tourists who decide to “visit again”.

Yi Geng,1 Yu-Jie Sun,1 Hao Song,1 Qiu-Ju Miao,1 Yi-Fei Wang,1 Jin-Liang Qi,2 Xiu-Lian Xu,1 Jian-Fang Sun1 1Institute of Dermatology, Peking Union Medical College and Chinese Academy of Medical Sciences, Nanjing, 210042, People’s Republic of China; 2State Key Laboratory of Pharmaceutical Biotechnology, School of Life Sciences, Nanjing University, Nanjing, 210023, People’s Republic of ChinaCorrespondence: Xiu-Lian Xu; Jian-Fang Sun, Email xxlqjl@sina.com; fangmin5758@aliyun.comBackground: Skin cutaneous melanoma (SKCM) is the deadliest dermatology tumor. Ongoing researches have confirmed that the NOD-like receptors (NLRs) family are crucial in driving carcinogenesis. However, the function of NLRs signaling pathway-related genes in SKCM remains unclear.Objective: To establish and identify an NLRs-related prognostic signature and to explore its predictive power for heterogeneous immune response in SKCM patients.Methods: Establishment of the predictive signature using the NLRs-related genes by least absolute shrinkage and selection operator-Cox regression analysis (LASSO-COX algorithm). Through univariate and multivariate COX analyses, NLRs signature’s independent predictive effectiveness was proven. CIBERSORT examined the comparative infiltration ratios of 22 distinct types of immune cells. RT-qPCR and immunohistochemistry implemented expression validation for critical NLRs-related prognostic genes in clinical samples.Results: The prognostic signature, including 7 genes, was obtained by the LASSO-Cox algorithm. In TCGA and validation cohorts, SKCM patients with higher risk scores had remarkably poorer overall survival. The independent predictive role of this signature was confirmed by multivariate Cox analysis. Additionally, a graphic nomogram demonstrated that the risk score of the NLRs signature has high predictive accuracy. SKCM patients in the low-risk group revealed a distinct immune microenvironment characterized by the significantly activated inflammatory response, interferon-α/γ response, and complement pathways. Indeed, several anti-tumor immune cell types were significantly accumulated in the low-risk group, including M1 macrophage, CD8 T cell, and activated NK cell. It is worth noting that our NLRs prognostic signature could serve as one of the promising biomarkers for predicting response rates to immune checkpoint blockade (ICB) therapy. Furthermore, the results of expression validation (RT-qPCR and IHC) were consistent with the previous analysis.Conclusion: A promising NLRs signature with excellent predictive efficacy for SKCM was developed.Keywords: cutaneous melanoma, NLR proteins, immune, survival analysis, prognosis

Many recent efforts aim to augment language models with relevant information retrieved from a database at test time. We avoid the need for prompt engineering by directly fine-tuning the model on data retrieved at test time using its standard training setup. For this purpose, we build a large-scale distributed nearest neighbor index based on text embeddings of the Pile dataset. Given a query to a language model, our system retrieves the neighbors of the query and fine-tunes the model on the text data corresponding to those neighbors. Surprisingly, retrieving and training on as few as 20 neighbors, each for only one gradient iteration, drastically improves performance across more than twenty language modeling tasks in the Pile benchmark. For example, test-time training significantly narrows the performance gap between a small GPT2 model and a GPTNeo model, more than ten times larger, that was specifically trained to convergence on the Pile. Sufficient index quality and size, however, are important. Our work establishes a valuable first baseline for implementing test-time training in the context of large language models, opening the door to numerous promising research avenues., Corrected Figure 8. Code repository here: https://github.com/socialfoundations/tttlm

Ranunculus sceleratus (family: Ranunculaceae) is a medicinally and economically important plant; however, gaps in taxonomic and species identification limit its practical applicability. This study aimed to sequence the chloroplast genome of R. sceleratus from Republic of Korea. Chloroplast sequences were compared and analyzed among Ranunculus species. The chloroplast genome was assembled from Illumina HiSeq 2500 sequencing raw data. The genome was 156,329 bp and had a typical quadripartite structure comprising a small single-copy region, a large single-copy region, and two inverted repeats. Fifty-three simple sequence repeats were identified in the four quadrant structural regions. The region between the ndhC and trnV-UAC genes could be useful as a genetic marker to distinguish between R. sceleratus populations from Republic of Korea and China. The Ranunculus species formed a single lineage. To differentiate between Ranunculus species, we identified 16 hotspot regions and confirmed their potential using specific barcodes based on phylogenetic tree and BLAST-based analyses. The ndhE, ndhF, rpl23, atpF, rps4, and rpoA genes had a high posterior probability of codon sites in positive selection, while the amino acid site varied between Ranunculus species and other genera. Comparison of the Ranunculus genomes provides useful information regarding species identification and evolution that could guide future phylogenetic analyses.

Emerging technologies have led to significant advancements in vehicle technology, particularly in the development and commercialization of autonomous vehicles. Self-driving technology aims to reduce collisions, save energy, and improve traffic conditions, and major car manufacturers and technology companies have invested nearly $80 billion in research and development. Experts predict that by 2030, self-driving cars will be reliable and affordable enough to replace most traditional vehicles. The market for autonomous driving is expected to reach $800 billion by 2035 and $7 trillion by 2050, potentially saving 585,000 lives between 2035 and 2045. The future of mobility will bring significant changes to the environment, with smarter and more connected roads, towns, and cities. Most people will likely experience autonomous driving through privately rented or shared ride vehicles. In response to this trend, governments and industry leaders have proposed selfdriving bus solutions. Supported by the government, Taiwanese firms have successfully developed a selfdriving bus by integrating automotive parts and systems and using artificial intelligence and deep learning for human-like vision and driving behavior. The self-driving bus has been verified through field tests, and its platform has been commercialized.

BackgroundTumor infiltrating lymphocytes (TILs), the main component in the tumor microenvironment, play a critical role in the antitumor immune response. Few studies have developed a prognostic model based on TILs in osteosarcoma.MethodsScRNA-seq data was obtained from our previous research and bulk RNA transcriptome data was from TARGET database. WGCNA was used to obtain the immune-related gene modules. Subsequently, we applied LASSO regression analysis and SVM algorithm to construct a prognostic model based on TILs marker genes. What’s more, the prognostic model was verified by external datasets and experiment in vitro. ResultsEleven cell clusters and 2044 TILs marker genes were identified. WGCNA results showed that 545 TILs marker genes were the most strongly related with immune. Subsequently, a risk model including 5 genes was developed. We found that the survival rate was higher in the low-risk group and the risk model could be used as an independent prognostic factor. Meanwhile, high-risk patients had a lower abundance of immune cell infiltration and many immune checkpoint genes were highly expressed in the low-risk group. The prognostic model was also demonstrated to be a good predictive capacity in external datasets. The result of RT-qPCR indicated that these 5 genes have differential expression which accorded with the predicting outcomes.ConclusionsThis study developed a new molecular signature based on TILs marker genes, which is very effective in predicting OS prognosis and immunotherapy response.

The study aims to investigate the role of the lipid mediator resolvin D1 (RvD1) in bacterial keratitis in a murine model.The effect of RvD1 on Pseudomonas aeruginosa-stimulated human corneal epithelial cells (HCECs) and mouse macrophages and dendritic cells (DCs) was assessed. C57BL/6 mouse corneas were abraded and treated with RvD1 after stimulation with P. aeruginosa, following which cytokine production level in the cornea and drainage lymph nodes was compared with that in controls. Corneal opacity and thickness were assessed using anterior segment photographs, and optical coherence tomography and corneal infiltrates were analyzed using immunohistochemistry for neutrophils.RvD1 significantly inhibited pro-inflammatory cytokine production in HCECs, mouse macrophages, and DCs. Corneal opacity and corneal thickness were reduced, and the development of corneal infiltrates, specifically neutrophils, was also significantly inhibited by RvD1 in response to stimulation with P. aeruginosa.RvD1 inhibits P. aeruginosa-induced corneal inflammation. This finding supports a potential therapeutic approach for patients with bacterial keratitis.

Extraintestinal pathogenic Escherichia coli (ExPEC) is a common anthropozoonotic pathogen that causes systemic infections. To establish infection, ExPEC must utilize essential nutrients including iron from the host. Transferrin is an important iron source for multiple bacteria. However, the mechanism by which ExPEC utilizes transferrin remains unclear. In this study, we found that iron-saturated holo-transferrin rather than iron-free apo-transferrin promoted the vitality of ExPEC in heat-inactivated human serum. The multifunctional protein Elongation factor Tu (EFTu) worked as a holo-transferrin binding protein. EFTu not only bound holo-transferrin rather than apo-transferrin but also released transferrin-related iron, with all domains of EFTu involved in holo-transferrin binding and iron release events. We also identified the surface location of EFTu on ExPEC. Overexpression of EFTu on the surface of nonpathogenic E. coli not only promoted the binding of bacteria to holo-transferrin but also facilitated the uptake of transferrin-related iron. More importantly, it significantly enhanced the survival of E. coli in heat-inactivated human serum, which was positively correlated with holo-transferrin but not apo-transferrin. Our research revealed a novel function of EFTu in binding holo-transferrin to promote iron uptake by bacteria, suggesting that EFTu was a potential virulence factor of ExPEC. In addition, our study provided research avenues into the iron acquisition and pathogenicity mechanisms of ExPEC.

Alt text plays a crucial role in providing individuals who rely on screen readers with a textual representation of visual content. However, dynamic and computer-generated virtual worlds comprised of scenes, 3D objects, and avatars in virtual reality (VR) remain mostly inaccessible. Equally immersive experiences are harder to guarantee, as conventional alt text may not provide rich enough descriptions for screen reader users. Additionally, engaging with virtual worlds only through the audio channel poses risks of auditory overload. To overcome the limitations of conventional alt text and reduce the experience gap between screen reader users and non-screen reader users in VR, we propose multisensory alt attributes. In this position paper, we explore the novel design space of using alt text (auditory channel) and alt sense (haptic channel) to convey the rich visual appearance of the VR world and consider the communication of social cues, and the support of navigation.

The work is devoted to the analysis of the choice of plasma-forming media for plasma torches used in complexes for plasma cutting of steel parts for shipbuilding problems. During plasma cutting, a heat-affected zone is formed with areas of structural changes and a cast layer on the cut surface. The length of the cast section is determined by the interaction of the plasma jet with the molten metal and its oxides during the cutting process. Minimizing the depth of the cast area reduces the roughness of the cut surface and the degree of unfavorable metallurgical changes in the properties of the parent metal at the cut edges. The effect of various plasma-forming gases on the quality of plasma cutting is analyzed: air, nitrogen, 80% nitrogen + 20% hydrogen, 50% argon + 50% oxygen, oxygen. It has been established that the highest saturation of the cast layer with nitrogen (up to 0.2-0.5%) is observed when air is used. The negative effect of nitrogen saturation of the cut edges is manifested in a significant increase in their hardness, as well as in the formation of pores in the welds obtained during subsequent welding of the cut parts. In the presence of hydrogen, the absorption of nitrogen by steel slows down. The rate of nitrogen desorption increases with an increase in the concentration of hydrogen in the gaseous medium to the power of 3/2. The binding of nitrogen in the plasma jet is positively affected by hydrocarbons. In addition, plasma-forming mixtures of air with hydrocarbons increase the energy efficiency of plasma cutting.

BackgroundThe assessment and management of neonatal pain is crucial for the development and wellbeing of vulnerable infants. Specifically, neonatal pain is associated with adverse health outcomes but is often under-identified and therefore under-treated. Neonatal stress may be misinterpreted as pain and may therefore be treated inappropriately. The assessment of neonatal pain is complicated by the non-verbal status of patients, age-dependent variation in pain responses, limited education on identifying pain in premature infants, and the clinical utility of existing tools.ObjectiveWe review research surrounding neonatal pain assessment scales currently in use to assess neonatal pain in the neonatal intensive care unit.MethodsWe performed a systematic review of original research using PRISMA guidelines for literature published between 2016 and 2021 using the key words “neonatal pain assessment” in the databases Web of Science, PubMed, and CINAHL. Fifteen articles remained after review, duplicate, irrelevant, or low-quality articles were eliminated.ResultsWe found research evaluating 13 neonatal pain scales. Important measurement categories include behavioral parameters, physiological parameters, continuous pain, acute pain, chronic pain, and the ability to distinguish between pain and stress. Provider education, inter-rater reliability and ease of use are important factors that contribute to an assessment tool's success. Each scale studied had strengths and limitations that aided or hindered its use for measuring neonatal pain in the neonatal intensive care unit, but no scale excelled in all areas identified as important for reliably identifying and measuring pain in this vulnerable population.ConclusionA more comprehensive neonatal pain assessment tool and more provider education on differences in pain signals in premature neonates may be needed to increase the clinical utility of pain scales that address the different aspects of neonatal pain.

IntroductionThe purpose of this paper is to effectively and accurately identify weed species in crop fields in complex environments. There are many kinds of weeds in the detection area, which are densely distributed.MethodsThe paper proposes the use of local variance pre-processing method for background segmentation and data enhancement, which effectively removes the complex background and redundant information from the data, and prevents the experiment from overfitting, which can improve the accuracy rate significantly. Then, based on the optimization improvement of DenseNet network, Efficient Channel Attention (ECA) mechanism is introduced after the convolutional layer to increase the weight of important features, strengthen the weed features and suppress the background features.ResultsUsing the processed images to train the model, the accuracy rate reaches 97.98%, which is a great improvement, and the comprehensive performance is higher than that of DenseNet, VGGNet-16, VGGNet-19, ResNet-50, DANet, DNANet, and U-Net models.DiscussionThe experimental data show that the model and method we designed are well suited to solve the problem of accurate identification of crop and weed species in complex environments, laying a solid technical foundation for the development of intelligent weeding robots.

Additional file 1: Table S1. Detailed definition of a healthy lifestyle. Table S2. HEI–2015 Components and Scoring Standards.

In recent years, there has been an increased popularity in image and speech generation using diffusion models. However, directly generating music waveforms from free-form text prompts is still under-explored. In this paper, we propose the first text-to-waveform music generation model that can receive arbitrary texts using diffusion models. We incorporate the free-form textual prompt as the condition to guide the waveform generation process of diffusion models. To solve the problem of lacking such text-music parallel data, we collect a dataset of text-music pairs from the Internet with weak supervision. Besides, we compare the effect of two prompt formats of conditioning texts (music tags and free-form texts) and prove the superior performance of our method in terms of text-music relevance. We further demonstrate that our generated music in the waveform domain outperforms previous works by a large margin in terms of diversity, quality, and text-music relevance.

Additional file 1: Supplementary Table 1. Multivariable negative binomial regression analyses for the number of emergency home visits and hospitalizations. Supplementary Table 2. Multivariable logistic regression analyses for emergency home visits and hospitalizations, excluding the dead during follow-up. Supplementary Table 3. Multivariable logistic regression analyses for emergency home visits and hospitalizations for patients without cancer diagnosis.

Supplementary Material 2

The double rod pendulum is a well known classic chaotic system, so its quantum version is an ideal laboratory to test various diagnosis for quantum chaos. We quantise this system canonically and calculate its lowest $10^4$ eigenvalues and eigenstate wave functions with at least $10^{-4}$ relative precision by the spectral analysis method. With these eigenvalues and eigenstate wave functions, we calculate and examine the three popular diagnosis on quantum chaos. On the NNSD diagnosis, we find that, either the GOE feature of NNSD is not a necessary condition for a quantum system to be chaotic at classic limit, or the double rod pendulum is not strong chaotic at the classic level. On the OTOC diagnosis, we observed that the early time exponential growth and late time constance approaching feature of OTOC is well conformed by the double rod pendulum. On the CC diagnosis, the status is similar with NNSD. Its linear growth feature at long time limit is either not a good diagnosis for a quantum system to be chaotic at classic limit or the double rod pendulum is not a strong chaotic system at classic levels., Comment: 9 pages, 9 figures

Heterogeneous Graph Neural Networks (HGNNs) have gained significant popularity in various heterogeneous graph learning tasks. However, most HGNNs rely on spatial domain-based message passing and attention modules for information propagation and aggregation. These spatial-based HGNNs neglect the utilization of spectral graph convolutions, which are the foundation of Graph Convolutional Networks (GCN) on homogeneous graphs. Inspired by the effectiveness and scalability of spectral-based GNNs on homogeneous graphs, this paper explores the extension of spectral-based GNNs to heterogeneous graphs. We propose PSHGCN, a novel heterogeneous convolutional network based on positive noncommutative polynomials. PSHGCN provides a simple yet effective approach for learning spectral graph convolutions on heterogeneous graphs. Moreover, we demonstrate the rationale of PSHGCN in graph optimization. We conducted an extensive experimental study to show that PSHGCN can learn diverse spectral heterogeneous graph convolutions and achieve superior performance in node classification tasks. Our code is available at https://github.com/ivam-he/PSHGCN., Comment: 10 pages

Compressed sensing magnetic resonance imaging (CS-MRI) seeks to recover visual information from subsampled measurements for diagnostic tasks. Traditional CS-MRI methods often separately address measurement subsampling, image reconstruction, and task prediction, resulting in suboptimal end-to-end performance. In this work, we propose TACKLE as a unified framework for designing CS-MRI systems tailored to specific tasks. Leveraging recent co-design techniques, TACKLE jointly optimizes subsampling, reconstruction, and prediction strategies to enhance the performance on the downstream task. Our results on multiple public MRI datasets show that the proposed framework achieves improved performance on various tasks over traditional CS-MRI methods. We also evaluate the generalization ability of TACKLE by experimentally collecting a new dataset using different acquisition setups from the training data. Without additional fine-tuning, TACKLE functions robustly and leads to both numerical and visual improvements.

Additional file 4: Fig. S1. Evaluation metrics for harmonization. (A) Batch correction performance evaluation by plots of kBET metrics and (B). Silhouette coefficients. For our demo dataset, LIGER was identified as the best batch correction method. Fig. S2. Percentage of top 50 features. Fig. S3. Number of cells with UMI > 0 for each gene (left) and the percentage of cells with UMI > 0 for each gene (right).

Prior work has established test-time training (TTT) as a general framework to further improve a trained model at test time. Before making a prediction on each test instance, the model is trained on the same instance using a self-supervised task, such as image reconstruction with masked autoencoders. We extend TTT to the streaming setting, where multiple test instances - video frames in our case - arrive in temporal order. Our extension is online TTT: The current model is initialized from the previous model, then trained on the current frame and a small window of frames immediately before. Online TTT significantly outperforms the fixed-model baseline for four tasks, on three real-world datasets. The relative improvement is 45% and 66% for instance and panoptic segmentation. Surprisingly, online TTT also outperforms its offline variant that accesses more information, training on all frames from the entire test video regardless of temporal order. This differs from previous findings using synthetic videos. We conceptualize locality as the advantage of online over offline TTT. We analyze the role of locality with ablations and a theory based on bias-variance trade-off., Comment: Project website with videos, dataset and code: https://video-ttt.github.io/

Additional file 3: File S1. The top-5 motifs of FaCINs for each chromosome.

Additional file 2. Selected illustrative quotes from student and staff focus groups.

Supplementary Material 1

Additional file 4 .

Additional file 1: Fig. S1. AUROC of DSB-GNN on each chromosome for NHEK cell line and performance comparison. Fig. S2. The distribution of the importance score of all node features and the top 10 k-mer features. Fig. S3. The comparison between the numbers of direct chromatin interactions and neck interactions for a prediction site on average. Fig. S4. Details for FaCIN’s bottleneck pattern. Fig. S5. An illustration for betweenness centrality and the comparison between neck neighbours and other neighbours in terms of betweenness centrality. Fig. S6. Results of subgraph search on FaCIN and random graph. Fig. S7. A schematic comparison between bottleneck and cycle patterns. Fig. S8. The 1D genomic length spanned by FaCIN’s interactions for 1-hop and 2-hop interactions. Fig. S9. The description of hypergeometric test for neck interactions and loop interactions.Fig. S10. Neck interactions in TAD boundaries. Fig. S11. The average length and intensity of neck interactions at DSB and non-DSB sites. Fig. S12. Percentage for neck neighbours of DSB prediction sites. Fig. S13. Statistics of the raw Hi-C contact counts. Fig. S14. A schematic for GNNExplainer masking approach to identify important edges.

Additional file 1: Fig. S1 a) Western blot analysis of USP1 expression after TAGLN overexpression. b) Duolink PLA demonstrates that USP1 interacts with TAGLN or ZEB1. c) Statistical analysis of Western blot results. d) Western blot analysis of ZEB1 protein expression after the co-knockdown of TAGLN and USP1. Fig. S2 SJB3-019A inhibited the deubiquitination of ZEB1 by USP1. Fig. S3 a) TIMP2 and NFKBIA mRNA levels quantified by q-PCR. b) Purified GST-tagged TAGLN was incubated with His-tagged USP1 in the presence or absence of the indicated amounts of Zerumbone. The interaction between TAGLN and USP1 was visualized using immunoblots. Fig. S4 Western blot analyses of ZEB1 and USP1 in the skin of WT mice with Zer or NE-Zer treatment before UV irradiation.