A spatiotemporal atlas of mouse liver homeostasis and regeneration

The mechanism by which mammalian liver cell responses are coordinated during tissue homeostasis and perturbation is poorly understood, representing a major obstacle in our understanding of many diseases. This knowledge gap is caused by the difficulty involved with studying multiple cell types in different states and locations, particularly when these are transient. We have combined Stereo-seq (spatiotemporal enhanced resolution omics-sequencing) with single-cell transcriptomic profiling of 473,290 cells to generate a high-definition spatiotemporal atlas of mouse liver homeostasis and regeneration at the whole-lobe scale. Our integrative study dissects in detail the molecular gradients controlling liver cell function, systematically defining how gene networks are dynamically modulated through intercellular communication to promote regeneration. Among other important regulators, we identified the transcriptional cofactor TBL1XR1 as a rheostat linking inflammation to Wnt/β-catenin signaling for facilitating hepatocyte proliferation. Our data and analytical pipelines lay the foundation for future high-definition tissue-scale atlases of organ physiology and malfunction.

A spatiotemporal atlas of cholestatic injury and repair in mice

Cholestatic liver injuries, characterized by regional damage around the bile ductular region, lack curative therapies and cause considerable mortality. Here we generated a high-definition spatiotemporal atlas of gene expression during cholestatic injury and repair in mice by integrating spatial enhanced resolution omics sequencing and single-cell transcriptomics. Spatiotemporal analyses revealed a key role of cholangiocyte-driven signaling correlating with the periportal damage-repair response. Cholangiocytes express genes related to recruitment and differentiation of lipid-associated macrophages, which generate feedback signals enhancing ductular reaction. Moreover, cholangiocytes express high TGFβ in association with the conversion of liver progenitor-like cells into cholangiocytes during injury and the dampened proliferation of periportal hepatocytes during recovery. Notably, Atoh8 restricts hepatocyte proliferation during 3,5-diethoxycarbonyl-1,4-dihydro-collidin damage and is quickly downregulated after injury withdrawal, allowing hepatocytes to respond to growth signals. Our findings lay a keystone for in-depth studies of cellular dynamics and molecular mechanisms of cholestatic injuries, which may further develop into therapies for cholangiopathies.

Biomaterial-based scaffolds for direct in situ programming of tumor-infiltrating T lymphocytes

Adoptive cell therapy with tumor-infiltrating T cells (TILs) has generated exciting clinical trial results for the treatment of unresectable solid tumors. However, solid tumors remain difficult targets for adoptively transferred T cells, due in part to poor migration of TILs to the tumor, physical barriers to infiltration, and active suppression of TILs by the tumor. Furthermore, a highly skilled team is required to obtain tumor tissue, isolate and expand the TILs ex vivo, and reinfuse them into the patient, which drives up costs and limits patient access. Here, we describe a cell-free polymer implant designed to recruit, genetically reprogram and expand host T cells at tumor lesions in situ. Importantly, the scaffold can be fabricated on a large scale and is stable to lyophilization. Using a mouse breast cancer model, we show that the implants quickly and efficiently amass cancer-specific host lymphocytes at the tumor site in quantities sufficient to bring about long-term tumor regression. Given that surgical care is the mainstay of cancer treatment for many patients, this technology could be easily implemented in a clinical setting as an add-on to surgery for solid tumors. Furthermore, the approach could be broadened to recruit and genetically reprogram other therapeutically desirable host cells, such as macrophages, natural killer cells or dendritic cells, potentially boosting the antitumor effectiveness of the implant even more.

A potential research target for cardiac rehabilitation: brain-derived neurotrophic factor

Cardiovascular diseases pose a major threat to human life, functional activity, and quality of life. Once the disease is present, patients can experience varying degrees of problems or limitations on three levels: physical, psychological, and social. Patients with cardiovascular disease are always at risk for adverse cardiac events, decreased physical activity, psychoemotional disturbances, and limited social participation due to their varying pathologies. Therefore, personalized cardiac rehabilitation is of great significance in improving patients' physical and mental functions, controlling disease progression, and preventing deterioration. There is a consensus on the benefits of cardiac rehabilitation in improving patients' quality of life, enhancing functional activity, and reducing mortality. As an important part of cardiac rehabilitation, Exercise plays an irreplaceable role. Aerobic exercise, resistance training, flexibility training, and other forms of exercise are recommended by many experts. Improvements in exercise tolerance, lipid metabolism, cardiac function, and psychological aspects of the patients were evident with appropriate exercise interventions based on a comprehensive assessment. Further studies have found that brain-derived neurotrophic factor may be an important mediator of exercise's ability to improve cardiovascular health. Brain-derived neurotrophic factor exerts multiple biological effects on the cardiovascular system. This article provides another perspective on the cardiac effects of exercise and further looks at the prospects for the use of brain-derived neurotrophic factor in cardiac rehabilitation. Meanwhile, the new idea that brain-derived neurotrophic factor is a key mediator connecting the brain-cardiac axis is proposed in light of the current research progress, to provide new ideas for clinical rehabilitation and scientific research.

MS²-GNN: Exploring GNN-Based Multimodal Fusion Network for Depression Detection

Major depressive disorder (MDD) is one of the most common and severe mental illnesses, posing a huge burden on society and families. Recently, some multimodal methods have been proposed to learn a multimodal embedding for MDD detection and achieved promising performance. However, these methods ignore the heterogeneity/homogeneity among various modalities. Besides, earlier attempts ignore interclass separability and intraclass compactness. Inspired by the above observations, we propose a graph neural network (GNN)-based multimodal fusion strategy named modal-shared modal-specific GNN, which investigates the heterogeneity/homogeneity among various psychophysiological modalities as well as explores the potential relationship between subjects. Specifically, we develop a modal-shared and modal-specific GNN architecture to extract the inter/intramodal characteristics. Furthermore, a reconstruction network is employed to ensure fidelity within the individual modality. Moreover, we impose an attention mechanism on various embeddings to obtain a multimodal compact representation for the subsequent MDD detection task. We conduct extensive experiments on two public depression datasets and the favorable results demonstrate the effectiveness of the proposed algorithm.

Effects of remittances on life expectancy and under-five mortality in sub-Saharan Africa: Evidence using Generalized Method of Moments analysis

The study examined the relationship between financial remittances and health outcomes in 45 sub-Saharan African countries (SSA) using data obtained from the World Development Indicator (WDI) over the period 1990 to 2021. Because of the issue of endogeneity, the System Generalized Method of Moments (SGMM) was adopted to analyze the impact of remittances on life expectancy and infant mortality respectively. The results showed that contrary to expectations, remittances did not significantly improve life expectancy and infant mortality rate in SSA. The life expectancy in the previous year, has a statically significant impact on life expectancy at birth for the current year. Also, the lagged value of infant mortality rate significantly increased under five mortality. Therefore, the study recommends that governments in SSA sub-region should evolve policies aimed at guiding recipients of remittances towards effective utilization with a view to improving social welfare and health outcomes.

Few-Shot Partial Multi-View Learning

It is often the case that data are with multiple views in real-world applications. Fully exploring the information of each view is significant for making data more representative. However, due to various limitations and failures in data collection and pre-processing, it is inevitable for real data to suffer from view missing and data scarcity. The coexistence of these two issues makes it more challenging to achieve the pattern classification task. Currently, to our best knowledge, few appropriate methods can well-handle these two issues simultaneously. Aiming to draw more attention from the community to this challenge, we propose a new task in this paper, called few-shot partial multi-view learning, which focuses on overcoming the negative impact of the view-missing issue in the low-data regime. The challenges of this task are twofold: (i) it is difficult to overcome the impact of data scarcity under the interference of missing views; (ii) the limited number of data exacerbates information scarcity, thus making it harder to address the view-missing issue in turn. To address these challenges, we propose a new unified Gaussian dense-anchoring method. The unified dense anchors are learned for the limited partial multi-view data, thereby anchoring them into a unified dense representation space where the influence of data scarcity and view missing can be alleviated. We conduct extensive experiments to evaluate our method. The results on Cub-googlenet-doc2vec, Handwritten, Caltech102, Scene15, Animal, ORL, tieredImagenet, and Birds-200-2011 datasets validate its effectiveness. The codes will be released at https://github.com/zhouyuan888888/UGDA.

A scATAC-seq atlas of chromatin accessibility in axolotl brain regions

Axolotl (Ambystoma mexicanum) is an excellent model for investigating regeneration, the interaction between regenerative and developmental processes, comparative genomics, and evolution. The brain, which serves as the material basis of consciousness, learning, memory, and behavior, is the most complex and advanced organ in axolotl. The modulation of transcription factors is a crucial aspect in determining the function of diverse regions within the brain. There is, however, no comprehensive understanding of the gene regulatory network of axolotl brain regions. Here, we utilized single-cell ATAC sequencing to generate the chromatin accessibility landscapes of 81,199 cells from the olfactory bulb, telencephalon, diencephalon and mesencephalon, hypothalamus and pituitary, and the rhombencephalon. Based on these data, we identified key transcription factors specific to distinct cell types and compared cell type functions across brain regions. Our results provide a foundation for comprehensive analysis of gene regulatory programs, which are valuable for future studies of axolotl brain development, regeneration, and evolution, as well as on the mechanisms underlying cell-type diversity in vertebrate brains.

Preparation and Characterization of Bio-Based PLA/PEG/g-C3N4 Low-Temperature Composite Phase Change Energy Storage Materials

As energy and environmental issues become more prominent, people must find sustainable, green development paths. Bio-based polymeric phase change energy storage materials provide solutions to cope with these problems. Therefore, in this paper, a fully degradable polyethylene glycol (PEG20000)/polylactic acid (PLA)/g-C₃N₄ composite phase change energy storage material (CPCM) was obtained by confinement. The CPCM was characterized by FTIR and SEM for compatibility, XRD and nanoindentation for mechanical properties and DSC, LFA, and TG for thermal properties. The results showed that the CPCM was physical co-mingling; when PLA: PEG: g-C₃N₄ was 6:3:1, the consistency was good. PEG destroys the crystallization of PLA and causes the hardness to decrease. When PLA: PEG: g-C₃N₄ was 6: 3: 1, it had a maximum hardness of 0.137 GPa. The CPCM had a high latent enthalpy, and endothermic and exothermic enthalpies of 106.1 kJ/kg and 80.05 kJ/kg for the PLA: PEG: g-C₃N₄ of 3: 6: 1. The CPCM showed an increased thermal conductivity compared to PLA, reaching 0.30 W/(m·K),0.32 W/(m·K) when PLA: PEG: g-C₃N₄ was 6: 3: 1 and when PLA: PEG: g-C₃N₄ was 3: 6: 1, respectively. Additionally, the CPCM was stable within 250 °C, indicating a wide appliable temperature range. The CPCM can be applied to solar thermal power generation, transportation, and building construction.

Single-cell chromatin accessibility profiling of cell-state-specific gene regulatory programs during mouse organogenesis

In mammals, early organogenesis begins soon after gastrulation, accompanied by specification of various type of progenitor/precusor cells. In order to reveal dynamic chromatin landscape of precursor cells and decipher the underlying molecular mechanism driving early mouse organogenesis, we performed single-cell ATAC-seq of E8.5-E10.5 mouse embryos. We profiled a total of 101,599 single cells and identified 41 specific cell types at these stages. Besides, by performing integrated analysis of scATAC-seq and public scRNA-seq data, we identified the critical cis-regulatory elements and key transcription factors which drving development of spinal cord and somitogenesis. Furthermore, we intersected accessible peaks with human diseases/traits-related loci and found potential clinical associated single nucleotide variants (SNPs). Overall, our work provides a fundamental source for understanding cell fate determination and revealing the underlying mechanism during postimplantation embryonic development, and expand our knowledge of pathology for human developmental malformations.

A novel computational method enables RNA editome profiling during human hematopoiesis from scRNA-seq data

RNA editing is a post-transcriptional modification with a cell-specific manner and important biological implications. Although single-cell RNA-seq (scRNA-seq) is an effective method for studying cellular heterogeneity, it is difficult to detect and study RNA editing events from scRNA-seq data because of the low sequencing coverage. To overcome this, we develop a computational method to systematically identify RNA editing sites of cell types from scRNA-seq data. To demonstrate its effectiveness, we apply it to scRNA-seq data of human hematopoietic stem/progenitor cells (HSPCs) with an annotated lineage differentiation relationship according to previous research and study the impacts of RNA editing on hematopoiesis. The dynamic editing patterns reveal the relevance of RNA editing on different HSPCs. For example, four microRNA (miRNA) target sites on 3' UTR of EIF2AK2 are edited across all HSPC populations, which may abolish the miRNA-mediated inhibition of EIF2AK2. Elevated EIF2AK2 may thus activate the integrated stress response (ISR) pathway to initiate global translational attenuation as a protective mechanism to maintain cellular homeostasis during HSPCs' differentiation. Besides, our findings also indicate that RNA editing plays an essential role in the coordination of lineage commitment and self-renewal of hematopoietic stem cells (HSCs). Taken together, we demonstrate the capacity of scRNA-seq data to exploit RNA editing events of cell types, and find that RNA editing may exert multiple modules of regulation in hematopoietic processes.

[Expression and clinical significance of plasma methylated SEPT 9 gene in patients with primary liver cancer]

Objective: To investigate the expression and clinical significance of plasma methylated SEPT9 (mSEPT9) gene in patients with primary liver cancer. Methods: 393 cases who visited our hospital from May 2016 to October 2018 were selected. Among them, 75 cases were in the primary liver cancer (PLC) group, 50 cases were in the liver cirrhosis (LC) group, and 268 cases were in the healthy control group (HC). The three groups' positive rates of mSEPT9 expression in the peripheral plasma were detected by the polymerase chain reaction (PCR) fluorescent probe method. The correlational clinical features of liver cancer were analyzed. At the same time, the electrochemiluminescence detection method was used to compare the AFP positive rate. Statistical analysis was conducted using chi-square tests or continuity-corrected chi-square tests. Results: 367 cases actually had valid samples. There were 64, 42, and 64 cases in the liver cancer group, cirrhosis group, and healthy control group, respectively. Among them, 34 cases of liver cancer were verified from pathological tissues. The positive rate of plasma mSEPT9 was significantly higher in the liver cancer group than that in the liver cirrhosis and healthy control groups [76.6% (49/64), 35.7% (15/42), and 3.8% (10/261), respectively], and the differences were statistically significant (χ (2) = 176.017, P < 0.001). The sensitivity of plasma mSEPT9 detection (76.6%) was significantly better in liver cancer (76.6%) than that of AFP patients (54.7%), and the difference was statistically significant (χ (2) = 6.788, P < 0.01). Compared with the single detection, the sensitivity and specificity of plasma mSEPT9 combined with AFP were significantly improved (89.7% vs. 96.3%, respectively). Patients with liver cancer aged≥50 years, with clinical stage II or above, and those with pathological signs of moderate to low differentiation had higher levels of plasma mSEPT9 positive expression, and the differences were statistically significant (χ (2) = 6.41, 9.279, 6.332, P < 0.05). During the follow-up period, the survival time of liver cancer patients with positive plasma mSEPT9 expression was significantly shorter than that of those with negative expression (310 ± 26 days vs. 487 ± 59 days, respectively), with statistically significant differences (Log Rank P = 0.039). Conclusion: In China, the positive rate of plasma mSEPT9 detection in liver cancer patients is higher than that of AFP in relation to age, clinical stage, and degree of tissue differentiation; additionally, it has certain survival predictive values. As a result, detecting this gene has important clinical significance and potential clinical application value in the non-invasive diagnosis and prognosis assessment of patients with primary liver cancer.

Effect of antioxidant procyanidin b2 (pcb2) on ovine oocyte developmental potential in response to in vitro maturation (ivm) and vitrification stress

BACKGROUND

It was demonstrated that external stress, such as in vitro maturation (IVM) and vitrification process can induce significantly reduced development capacity in oocytes. Previous studies indicated that antioxidants play a pivotal part in the acquisition of adaptation in changed conditions. At present, the role of the natural potent antioxidant PCB2 in response to IVM and vitrification during ovine oocyte manipulation has not been explored.

OBJECTIVE

To investigate whether PCB2 treatment could improve the developmental potential of ovine oocytes under IVM and vitrification stimuli.

MATERIALS AND METHODS

The experiment was divided into two parts. Firstly, the effect of PCB2 on the development of oocytes during IVM was evaluated. Un-supplemented and 5 ug per mL PCB2-supplemented in the IVM solution were considered as control and experimental groups (C + 5 ug per mL PCB2). The polar body extrusion (PBE) rate, mitochondrial membrane potential (MMP), ATP, reactive oxygen species (ROS) levels and early apoptosis of oocytes were measured after IVM. Secondly, we further determine whether PCB2 could improve oocyte quality under vitrification stress. The survival rate, PBE rate and early apoptosis of oocytes were compared between fresh group, vitrified group and 5 ug per mL PCB2-supplemented in the IVM solution after vitrification (V + 5 ug per mL PCB2).

RESULTS

Compared to the control group, adding PCB2 significantly increased PBE rate (79.4% vs. 62.8%, P < 0.01) and MMP level (1.9 +/- 0.08 vs. 1.3 +/- 0.04, P < 0.01), and decreased ROS level (47.1 +/- 6.3 vs. 145.3 +/- 8.9, P < 0.01). However, there was no significant difference in ATP content and early apoptosis. Compared to the fresh group, vitrification significantly reduced oocytes viability (43.0% vs. 90.8%, P < 0.01) as well as PBE rate (24.2% vs. 60.6%, P < 0.05). However, 5 ug per mL PCB2-supplemention during maturation had no effect on survival, PBE or early apoptosis in vitrified oocytes.

CONCLUSION

PCB2 could effectively antagonise the oxidative stress during IVM and promote oocyte development. DOI: 10.54680/fr23210110412.

Ru Single Atoms on One-Dimensional CF@g-C3N4 Hierarchy as Highly Stable Catalysts for Aqueous Levulinic Acid Hydrogenation

Herein, we report a stable catalyst with Ru single atoms anchored on a one-dimensional carbon fiber@graphitic carbon nitride hierarchy, by assembling wet wipes composed of fiber-derived carbon fiber (CF), melamine-derived graphitic carbon nitride (g-C3N4) and RuCl3 before NaBH4 reduction. The atomically dispersed Ru species (3.0 wt%) are tightly attached via N-coordination provided by exterior g-C3N4 nanosheets, and further stabilized by the interior mesoporous CF. The obtained CF@g-C3N4-Ru SAs catalyst can be cycled six times without notable leaching of Ru or loss of GVL yield in the acidic media. This catalyst is more stable than Ru nanoparticles supported on CF@g-C3N4, as well as Ru single atoms anchored on CF and g-C3N4, and proves to be one of the most efficient metal catalysts for aqueous LA hydrogenation to γ-valerolactone (GVL). The isolated Ru atoms by strong N-coordination, and their enhanced electron/mass transfer afforded by the one-dimensional hierarchy, can be responsible for the excellent durability of CF@g-C3N4-Ru SAs under harsh reaction conditions.

Chromosomal-level genome of velvet bean (Mucuna pruriens) provides resources for L-DOPA synthetic research and development

Mucuna pruriens, commonly called velvet bean, is the main natural source of levodopa (L-DOPA), which has been marketed as a psychoactive drug for the clinical management of Parkinson’s disease and dopamine-responsive dystonia. Although velvet bean is a very important plant species for food and pharmaceutical manufacturing, the lack of genetic and genomic information about this species severely hinders further molecular research thereon and biotechnological development. Here, we reported the first velvet bean genome, with a size of 500.49 Mb and 11 chromosomes encoding 28,010 proteins. Genomic comparison among legume species indicated that velvet bean speciated ∼29 Ma from soybean clade, without specific genome duplication. Importantly, we identified 21 polyphenol oxidase coding genes that catalyse l-tyrosine to L-DOPA in velvet bean, and two subfamilies showing tandem expansion on Chr3 and Chr7 after speciation. Interestingly, disease-resistant and anti-pathogen gene families were found contracted in velvet bean, which might be related to the expansion of polyphenol oxidase. Our study generated a high-quality genomic reference for velvet bean, an economically important agricultural and medicinal plant, and the newly reported L-DOPA biosynthetic genes could provide indispensable information for the biotechnological and sustainable development of an environment-friendly L-DOPA biosynthesis processing method.

[Factors analysis of worsening renal function in patients with acute right ventricular myocardial infarction during hospitalization]

Objective: To analyze the related factors of worsening renal function (WRF) in patients with acute right ventricular myocardial infarction (RVMI) during hospitalization. Methods: A total of 98 patients with acute RVMI admitted to the emergency comprehensive ward of Beijing Anzhen Hospital from August 2011 to January 2020 were enrolled in this cross-sectional study. According to the situation of WRF, the patients were divided into non-WRF group (76 cases) and WRF group (22 cases). WRF was defined as ≥0.3 mg/dL increase in serum creatinine level from baseline on day 6 of hospitalization (if hospital stay<6 days, it was at discharge). Baseline data, intravenous fluid infusion, diuretic and significant positive balance of patients' intake and output volume [any 24 h intakes and outputs ≥1 000 ml or any consecutive 72 h intakes and outputs ≥2 000 ml within 6 d of hospitalization (if hospitalization<6 d, it was from admission to discharge)] were obtained, and the differences of above indicators between the two groups were analyzed. Multiple logistic regression model was used to analyze the related factors of WRF. Results: The ages of patients in WRF group and non-WRF group were 60 (50, 68) and 63 (52, 72) years, and the male proportions were 63.6% (14 cases) and 76.3% (58 cases), respectively, and there was no significant difference (all P>0.05). The proportion of positive balance was 31.8% (7 cases) in WRF group, which was higher than 14.5% (11 cases) in non-WRF group (P=0.034). The rate of loop diuretic use in WRF group was 4.5% (1 case), lower than that in non-WRF group 10.5% (8 cases) (P=0.027). After adjusting for age, sex, baseline estimated glomerular filtration rate (eGFR), preoperative isoproterenol/temporary pacemaker/atropine use, significant positive balance of intake and output volume, and loop diuretic use, it was found that eGFR≥60 ml·min^-1·1.73 m^-2 and significant positive balance were associated with WRF, the OR (95%CI) were 0.71 (0.62-0.86) and 1.21 (1.02-1.43) (both P<0.05); After eliminating the variable of significant positive balance in the above model, loop diuretic use was found to be a correlation factor for WRF, with an OR (95%CI) of 0.89 (0.72-0.97) (P<0.05). Conclusions: Significant positive balance of intake and output volume during hospitalization in patients with acute RVMI is a risk factor for WRF on day 6 or at discharge. In the presence of a significant positive balance, loop diuretic use is a protective factor for WRF.

[Clinical phenotype analysis of 6 cases of TTC21B gene related nephronophthisis]

Objective: To analyze the clinical characteristics of 6 children with TTC21B-related nephronophthisis to provide reference for early clinical diagnosis. Methods: The general condition, clinical manifestations, laboratory tests and other clinical data of 6 children from 4 families diagnosed with nephronophthisis by genetic testing in Shanghai Children's Hospital from January 2015 to December 2020 were analyzed retrospectively. Results: A total of 6 children (3 males and 3 females) developed proteinuria and progressive renal dysfunction in early infancy. The onset age of proteinuria was 18 (6, 25) months. The age at the onset of renal impairment was 22 (10, 36) months. All 6 children progressed to end-stage renal disease (ESRD) within 10 (4, 65) months of onset. Five children had hypertension, 3 children with abnormal liver function, 2 children with visceral translocation and 1 child with growth retardation. The genetic results suggested that all children carried variations TTC21B gene p.C518R. Conclusions: Children with TTC21B gene p.C518R nephronophthisis had proteinuria and progressed to ESRD at the early stage of life. These nephronophthisis patients commonly presented with liver and renal dysfunction.

Differentiated Explanation of Deep Neural Networks With Skewed Distributions

Over the last decade, deep neural networks (DNNs) are regarded as black-box methods, and their decisions are criticized for the lack of explainability. Existing attempts based on local explanations offer each input a visual saliency map, where the supporting features that contribute to the decision are emphasized with high relevance scores. In this paper, we improve the saliency map based on differentiated explanations, of which the saliency map not only distinguishes the supporting features from backgrounds but also shows the different degrees of importance of the various parts within the supporting features. To do this, we propose to learn a differentiated relevance estimator called DRE, where a carefully-designed distribution controller is introduced to guide the relevance scores towards right-skewed distributions. DRE can be directly optimized under pure classification losses, enabling higher faithfulness of explanations and avoiding non-trivial hyper-parameter tuning. The experimental results on three real-world datasets demonstrate that our differentiated explanations significantly improve the faithfulness with high explainability. Our code and trained models are available at https://github.com/fuweijie/DRE.

Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays

Spatially resolved transcriptomic technologies are promising tools to study complex biological processes such as mammalian embryogenesis. However, the imbalance between resolution, gene capture, and field of view of current methodologies precludes their systematic application to analyze relatively large and three-dimensional mid- and late-gestation embryos. Here, we combined DNA nanoball (DNB)-patterned arrays and in situ RNA capture to create spatial enhanced resolution omics-sequencing (Stereo-seq). We applied Stereo-seq to generate the mouse organogenesis spatiotemporal transcriptomic atlas (MOSTA), which maps with single-cell resolution and high sensitivity the kinetics and directionality of transcriptional variation during mouse organogenesis. We used this information to gain insight into the molecular basis of spatial cell heterogeneity and cell fate specification in developing tissues such as the dorsal midbrain. Our panoramic atlas will facilitate in-depth investigation of longstanding questions concerning normal and abnormal mammalian development.

Hierarchical Prototype Refinement With Progressive Inter-Categorical Discrimination Maximization for Few-Shot Learning

Metric-based few-shot learning categorizes unseen query instances by measuring their distance to the categories appearing in the given support set. To facilitate distance measurement, prototypes are used to approximate the representations of categories. However, we find prototypical representations are generally not discriminative enough to represent the discrepancy of inter-categorical distribution of queries, thereby limiting the classification accuracy. To overcome this issue, we propose a new Progressive Hierarchical-Refinement (PHR) method, which effectively refines the discrimination of prototypes by conducting the Progressive Discrimination Maximization strategy based on the hierarchical feature representations. Specifically, we first encode supports and queries into the representation space of spatial level, global level, and semantic level. Then, the refining coefficients are constructed by exploring the metric information contained in these hierarchical embedding spaces simultaneously. Under the guidance of the refining coefficients, the meta-refining loss progressively maximizes the discrimination degree of inter-categorical prototypical representations. In addition, the refining vectors are adopted to further enhance the representations of prototypes. In this way, the metric-based classification can be more accurate. Our PHR method shows the competitive performance on the miniImagenet, CIFAR-FS, FC100, and CUB datasets. Moreover, PHR presents good compatibility. It can be incorporated with other few-shot learning models, making them more accurate.