Neighbor-specific gene expression revealed from physically interacting cells during mouse embryonic development.

Development of multicellular organisms is orchestrated by persistent cell-cell communication between neighboring partners. Direct interaction between different cell types can induce molecular signals that dictate lineage specification and cell fate decisions. Current single-cell RNA-seq technology cannot adequately analyze cell-cell contact-dependent gene expression, mainly due to the loss of spatial information. To overcome this obstacle and resolve cell-cell contact-specific gene expression during embryogenesis, we performed RNA sequencing of physically interacting cells (PIC-seq) and assessed them alongside similar single-cell transcriptomes derived from developing mouse embryos between embryonic day (E) 7.5 and E9.5. Analysis of the PIC-seq data identified gene expression signatures that were dependent on the presence of specific neighboring cell types. Our computational predictions, validated experimentally, demonstrated that neural progenitor (NP) cells upregulate Lhx5 and Nkx2-1 genes, when exclusively interacting with definitive endoderm (DE) cells. Moreover, there was a reciprocal impact on the transcriptome of DE cells, as they tend to upregulate Rax and Gsc when in contact with NP cells. Using individual cell transcriptome data, we formulated a means of computationally predicting the impact of one cell type on the transcriptome of its neighboring cell types. We have further developed a distinctive spatial-t-distributed stochastic neighboring embedding to display the pseudospatial distribution of cells in a 2-dimensional space. In summary, we describe an innovative approach to study contact-specific gene regulation during embryogenesis.

A PDX1 cistrome and single-cell transcriptome resource of the developing pancreas.

Pancreatic and duodenal homeobox 1 (PDX1) is crucial for pancreas organogenesis, yet the dynamic changes in PDX1 binding in human or mouse developing pancreas have not been examined. To address this knowledge gap, we performed PDX1 ChIP-seq and single-cell RNA-seq using fetal human pancreata. We integrated our datasets with published datasets and revealed the dynamics of PDX1 binding and potential cell lineage-specific PDX1-bound genes in the pancreas from fetal to adult stages. We identified a core set of developmentally conserved PDX1-bound genes that reveal the broad multifaceted role of PDX1 in pancreas development. Despite the well-known dramatic changes in PDX1 function and expression, we found that PDX1-bound genes are largely conserved from embryonic to adult stages. This points towards a dual role of PDX1 in regulating the expression of its targets at different ages, dependent on other functionally congruent or directly interacting partners. We also showed that PDX1 binding is largely conserved in mouse pancreas. Together, our study reveals PDX1 targets in the developing pancreas in vivo and provides an essential resource for future studies on pancreas development.

CellNeighborEX: deciphering neighbor-dependent gene expression from spatial transcriptomics data.

Cells have evolved their communication methods to sense their microenvironments and send biological signals. In addition to communication using ligands and receptors, cells use diverse channels including gap junctions to communicate with their immediate neighbors. Current approaches, however, cannot effectively capture the influence of various microenvironments. Here, we propose a novel approach to investigate cell neighbor-dependent gene expression (CellNeighborEX) in spatial transcriptomics (ST) data. To categorize cells based on their microenvironment, CellNeighborEX uses direct cell location or the mixture of transcriptome from multiple cells depending on ST technologies. For each cell type, CellNeighborEX identifies diverse gene sets associated with partnering cell types, providing further insight. We found that cells express different genes depending on their neighboring cell types in various tissues including mouse embryos, brain, and liver cancer. Those genes are associated with critical biological processes such as development or metastases. We further validated that gene expression is induced by neighboring partners via spatial visualization. The neighbor-dependent gene expression suggests new potential genes involved in cell-cell interactions beyond what ligand-receptor co-expression can discover.

Systemic approaches using single cell transcriptome reveal that C/EBPγ regulates autophagy under amino acid starved condition.

Autophagy, a catabolic process to remove unnecessary or dysfunctional organelles, is triggered by various signals including nutrient starvation. Depending on the types of the nutrient deficiency, diverse sensing mechanisms and signaling pathways orchestrate for transcriptional and epigenetic regulation of autophagy. However, our knowledge about nutrient type-specific transcriptional regulation during autophagy is limited. To understand nutrient type-dependent transcriptional mechanisms during autophagy, we performed single cell RNA sequencing (scRNAseq) in the mouse embryonic fibroblasts (MEFs) with or without glucose starvation (GS) as well as amino acid starvation (AAS). Trajectory analysis using scRNAseq identified sequential induction of potential transcriptional regulators for each condition. Gene regulatory rules inferred using TENET newly identified CCAAT/enhancer binding protein γ (C/EBPγ) as a regulator of autophagy in AAS, but not GS, condition, and knockdown experiment confirmed the TENET result. Cell biological and biochemical studies validated that activating transcription factor 4 (ATF4) is responsible for conferring specificity to C/EBPγ for the activation of autophagy genes under AAS, but not under GS condition. Together, our data identified C/EBPγ as a previously unidentified key regulator under AAS-induced autophagy.

SHARP: hyperfast and accurate processing of single-cell RNA-seq data via ensemble random projection.

To process large-scale single-cell RNA-sequencing (scRNA-seq) data effectively without excessive distortion during dimension reduction, we present SHARP, an ensemble random projection-based algorithm that is scalable to clustering 10 million cells. Comprehensive benchmarking tests on 17 public scRNA-seq data sets show that SHARP outperforms existing methods in terms of speed and accuracy. Particularly, for large-size data sets (more than 40,000 cells), SHARP runs faster than other competitors while maintaining high clustering accuracy and robustness. To the best of our knowledge, SHARP is the only R-based tool that is scalable to clustering scRNA-seq data with 10 million cells.

TENET: gene network reconstruction using transfer entropy reveals key regulatory factors from single cell transcriptomic data.

Accurate prediction of gene regulatory rules is important towards understanding of cellular processes. Existing computational algorithms devised for bulk transcriptomics typically require a large number of time points to infer gene regulatory networks (GRNs), are applicable for a small number of genes and fail to detect potential causal relationships effectively. Here, we propose a novel approach 'TENET' to reconstruct GRNs from single cell RNA sequencing (scRNAseq) datasets. Employing transfer entropy (TE) to measure the amount of causal relationships between genes, TENET predicts large-scale gene regulatory cascades/relationships from scRNAseq data. TENET showed better performance than other GRN reconstructors, in identifying key regulators from public datasets. Specifically from scRNAseq, TENET identified key transcriptional factors in embryonic stem cells (ESCs) and during direct cardiomyocytes reprogramming, where other predictors failed. We further demonstrate that known target genes have significantly higher TE values, and TENET predicted higher TE genes were more influenced by the perturbation of their regulator. Using TENET, we identified and validated that Nme2 is a culture condition specific stem cell factor. These results indicate that TENET is uniquely capable of identifying key regulators from scRNAseq data.

VeTra: a tool for trajectory inference based on RNA velocity.

MOTIVATION: Trajectory inference (TI) for single cell RNA sequencing (scRNAseq) data is a powerful approach to interpret dynamic cellular processes such as cell cycle and development. Still, however, accurate inference of trajectory is challenging. Recent development of RNA velocity provides an approach to visualize cell state transition without relying on prior knowledge. RESULTS: To perform TI and group cells based on RNA velocity we developed VeTra. By applying cosine similarity and merging weakly connected components, VeTra identifies cell groups from the direction of cell transition. Besides, VeTra suggests key regulators from the inferred trajectory. VeTra is a useful tool for TI and subsequent analysis. AVAILABILITY AND IMPLEMENTATION: The Vetra is available at https://github.com/wgzgithub/VeTra. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

FOXO1 Is a Key Mediator of Glucocorticoid-Induced Expression of Tristetraprolin in MDA-MB-231 Breast Cancer Cells.

The mRNA destabilizing factor tristetraprolin (TTP) functions as a tumor suppressor by down-regulating cancer-associated genes. TTP expression is significantly reduced in various cancers, which contributes to cancer processes. Enforced expression of TTP impairs tumorigenesis and abolishes maintenance of the malignant state, emphasizing the need to identify a TTP inducer in cancer cells. To search for novel candidate agents for inducing TTP in cancer cells, we screened a library containing 1019 natural compounds using MCF-7 breast cancer cells transfected with a reporter vector containing the TTP promoter upstream of the luciferase gene. We identified one molecule, of which the enantiomers are betamethasone 21-phosphate (BTM-21-P) and dexamethasone 21-phosphate (BTM-21-P), as a potent inducer of TTP in cancer cells. We confirmed that BTM-21-P, DXM-21-P, and dexamethasone (DXM) induced the expression of TTP in MDA-MB-231 cells in a glucocorticoid receptor (GR)-dependent manner. To identify potential pathways linking BTM-21-P and DXM-21-P to TTP induction, we performed an RNA sequencing-based transcriptome analysis of MDA-MB-231 cells at 3 h after treatment with these compounds. A heat map analysis of FPKM expression showed a similar expression pattern between cells treated with the two compounds. The KEGG pathway analysis results revealed that the upregulated DEGs were strongly associated with several pathways, including the Hippo signaling pathway, PI3K-Akt signaling pathway, FOXO signaling pathway, NF-κB signaling pathway, and p53 signaling pathway. Inhibition of the FOXO pathway using a FOXO1 inhibitor blocked the effects of BTM-21-P and DXM-21-P on the induction of TTP in MDA-MB-231 cells. We found that DXM enhanced the binding of FOXO1 to the TTP promoter in a GR-dependent manner. In conclusion, we identified a natural compound of which the enantiomers are DXM-21-P and BTM-21-P as a potent inducer of TTP in breast cancer cells. We also present new insights into the role of FOXO1 in the DXM-21-P- and BTM-21-P-induced expression of TTP in cancer cells.

Dietary Collagen Hydrolysates Ameliorate Furrowed and Parched Skin Caused by Photoaging in Hairless Mice.

Collagen hydrolysates have been suggested as a favorable antiaging modality in skin photoaged by persistent exposure to ultraviolet radiation (UV). The current study evaluated the beneficial effect of collagen hydrolysates (fsCH) extracted from Pangasius hypophthalmus fish skin on wrinkle formation and moisture preservation in dorsal skin of hairless mice challenged with UV-B. Inter-comparative experiments were conducted for anti-photoaging among fsCH, retinoic acid (RA), N-acetyl-D-glucosamine (NAG), and glycine-proline-hydroxyproline (GPH). Treating human HaCaT keratinocytes with 100-200 µg/mL fsCH reciprocally ameliorated the expression of aquaporin 3 (AQP3) and CD44 deranged by UV-B. The UV-B-induced deep furrows and skin thickening were improved in parched dorsal skin of mice supplemented with 206-412 mg/kg fsCH as well as RA and GPH. The UV-B irradiation enhanced collagen fiber loss in the dorsal dermis, which was attenuated by fsCH through enhancing procollagen conversion to collagen. The matrix metalloproteinase expression by UV-B in dorsal skin was diminished by fsCH, similar to RA and GPH, via blockade of collagen degradation. Supplementing fsCH to UV-B-irradiated mice decreased transepidermal water loss in dorsal skin with reduced AQP3 level and restored keratinocyte expression of filaggrin. The expression of hyaluronic acid synthase 2 and hyaluronidase 1 by UV-B was remarkably ameliorated with increased production of hyaluronic acid by treating fsCH to photoaged mice. Taken together, fsCH attenuated photoaging typical of deep wrinkles, epidermal thickening, and skin water loss, like NAG, RA, or GPH, through inhibiting collagen destruction and epidermal barrier impairment.

CellBIC: bimodality-based top-down clustering of single-cell RNA sequencing data reveals hierarchical structure of the cell type.

Single-cell RNA sequencing (scRNA-seq) is a powerful tool to study heterogeneity and dynamic changes in cell populations. Clustering scRNA-seq is essential in identifying new cell types and studying their characteristics. We develop CellBIC (single Cell BImodal Clustering) to cluster scRNA-seq data based on modality in the gene expression distribution. Compared with classical bottom-up approaches that rely on a distance metric, CellBIC performs hierarchical clustering in a top-down manner. CellBIC outperformed the bottom-up hierarchical clustering approach and other recently developed clustering algorithms while maintaining the hierarchical structure of cells. Importantly, CellBIC identifies type 2 diabetes and age specific ß cell signatures characterized by SIX3 and CDH2, respectively.

Bone Remodeling: Histone Modifications as Fate Determinants of Bone Cell Differentiation.

The bone tissue is a dynamic complex that constitutes of several interdependent systems and is continuously remodeled through the concerted actions of bone cells. Osteoblasts are mononucleated cells, derived from mesenchymal stem cells, responsible for bone formation. Osteoclasts are large multinucleated cells that differentiate from hematopoietic progenitors of the myeloid lineage and are responsible for bone resorption. The lineage-specific differentiation of bone cells requires an epigenetic regulation of gene expressions involving chromatin dynamics. The key step for understanding gene regulatory networks during bone cell development lies in characterizing the chromatin modifying enzymes responsible for reorganizing and potentiating particular chromatin structure. This review covers the histone-modifying enzymes involved in bone development, discusses the impact of enzymes on gene expression, and provides future directions and clinical significance in this area.

Homozygous deletions at 3p22, 5p14, 6q15, and 9p21 result in aberrant expression of tumor suppressor genes in gastric cancer.

Homozygous deletion is a frequent mutational mechanism of silencing tumor suppressor genes in cancer. Therefore, homozygous deletions have been analyzed for identification of tumor suppressor genes that can be utilized as biomarkers or therapeutic targets for cancer treatment. In this study, to elucidate potential tumor suppressor genes involved in gastric cancer (GC), we analyzed the entire set of large homozygous deletions in six human GC cell lines through genome- and transcriptome-wide approaches. We identified 51 genes in homozygous deletion regions of chromosomes and confirmed the deletion frequency in tumor tissues of 219 GC patients from The Cancer Genome Atlas database. We evaluated the effect of homozygous deletions on the mRNA level and found significantly affected genes in chromosome bands 9p21, 3p22, 5p14, and 6q15. Among the genes in 9p21, we investigated the potential tumor suppressive effect of KLHL9. We demonstrated that ectopic expression of KLHL9 inhibited cell proliferation and tumor formation in KLHL9-deficient SNU-16 cell line. In addition, we observed that homozygous focal deletions generated truncated transcripts of TGFBR2, CTNNA1, and STXBP5. Ectopic expression of two kinds of TGFBR2-reverse GADL1 fusion genes suppressed TGF-ß signaling, which may lead to the loss of sensitivity to TGF-ß tumor suppressive activity. In conclusion, our findings suggest that novel tumor suppressor genes that are aberrantly expressed through homozygous deletions may play important roles in gastric tumorigenesis.

Robustness and evolvability of the human signaling network.

Biological systems are known to be both robust and evolvable to internal and external perturbations, but what causes these apparently contradictory properties? We used Boolean network modeling and attractor landscape analysis to investigate the evolvability and robustness of the human signaling network. Our results show that the human signaling network can be divided into an evolvable core where perturbations change the attractor landscape in state space, and a robust neighbor where perturbations have no effect on the attractor landscape. Using chemical inhibition and overexpression of nodes, we validated that perturbations affect the evolvable core more strongly than the robust neighbor. We also found that the evolvable core has a distinct network structure, which is enriched in feedback loops, and features a higher degree of scale-freeness and longer path lengths connecting the nodes. In addition, the genes with high evolvability scores are associated with evolvability-related properties such as rapid evolvability, low species broadness, and immunity whereas the genes with high robustness scores are associated with robustness-related properties such as slow evolvability, high species broadness, and oncogenes. Intriguingly, US Food and Drug Administration-approved drug targets have high evolvability scores whereas experimental drug targets have high robustness scores.

The co-regulation mechanism of transcription factors in the human gene regulatory network.

The co-regulation of transcription factors (TFs) has been widely observed in various species. Why is such a co-regulation mechanism needed for transcriptional regulation? To answer this question, the following experiments and analyses were performed. First, examination of the human gene regulatory network (GRN) indicated that co-regulation was significantly enriched in the human GRN. Second, mathematical simulation of an artificial regulatory network showed that the co-regulation mechanism was related to the biphasic dose-response patterns of TFs. Third, the relationship between the co-regulation mechanism and the biphasic dose-response pattern was confirmed using microarray experiments examining different time points and different doses of the toxicant tetrachlorodibenzodioxin. Finally, two mathematical models were constructed to mimic highly co-regulated networks (HCNs) and little co-regulated networks (LCNs), and we found that HCNs were more robust to parameter perturbation than LCNs, whereas LCNs were faster in adaptation to environmental changes than HCNs.

A review on gene regulatory network reconstruction algorithms based on single cell RNA sequencing.

BACKGROUND: Understanding gene regulatory networks (GRNs) is essential for unraveling the molecular mechanisms governing cellular behavior. With the advent of high-throughput transcriptome measurement technology, researchers have aimed to reverse engineer the biological systems, extracting gene regulatory rules from their outputs, which represented by gene expression data. Bulk RNA sequencing, a widely used method for measuring gene expression, has been employed for GRN reconstruction. However, it falls short in capturing dynamic changes in gene expression at the level of individual cells since it averages gene expression across mixed cell populations. OBJECTIVE: In this review, we provide an overview of 15 GRN reconstruction tools and discuss their respective strengths and limitations, particularly in the context of single cell RNA sequencing (scRNA-seq). METHODS: Recent advancements in scRNA-seq break new ground of GRN reconstruction. They offer snapshots of the individual cell transcriptomes and capturing dynamic changes. We emphasize how these technological breakthroughs have enhanced GRN reconstruction. CONCLUSION: GRN reconstructors can be classified based on their requirement for cellular trajectory, which represents a dynamical cellular process including differentiation, aging, or disease progression. Benchmarking studies support the superiority of GRN reconstructors that do not require trajectory analysis in identifying regulator-target relationships. However, methods equipped with trajectory analysis demonstrate better performance in identifying key regulatory factors. In conclusion, researchers should select a suitable GRN reconstructor based on their specific research objectives.

Advances in single-cell omics and multiomics for high-resolution molecular profiling.

Single-cell omics technologies have revolutionized molecular profiling by providing high-resolution insights into cellular heterogeneity and complexity. Traditional bulk omics approaches average signals from heterogeneous cell populations, thereby obscuring important cellular nuances. Single-cell omics studies enable the analysis of individual cells and reveal diverse cell types, dynamic cellular states, and rare cell populations. These techniques offer unprecedented resolution and sensitivity, enabling researchers to unravel the molecular landscape of individual cells. Furthermore, the integration of multimodal omics data within a single cell provides a comprehensive and holistic view of cellular processes. By combining multiple omics dimensions, multimodal omics approaches can facilitate the elucidation of complex cellular interactions, regulatory networks, and molecular mechanisms. This integrative approach enhances our understanding of cellular systems, from development to disease. This review provides an overview of the recent advances in single-cell and multimodal omics for high-resolution molecular profiling. We discuss the principles and methodologies for representatives of each omics method, highlighting the strengths and limitations of the different techniques. In addition, we present case studies demonstrating the applications of single-cell and multimodal omics in various fields, including developmental biology, neurobiology, cancer research, immunology, and precision medicine.

DRG2 is required for surface localization of PD-L1 and the efficacy of anti-PD-1 therapy.

More than half of tumor patients with high PD-L1 expression do not respond to anti-PD-1/PD-L1 therapy, and the underlying mechanisms are yet to be clarified. Here we show that developmentally regulated GTP-binding protein 2 (DRG2) is required for response of PD-L1-expressing tumors to anti-PD-1 therapy. DRG2 depletion enhanced IFN-γ signaling and increased the PD-L1 level in melanoma cells. However, it inhibited recycling of endosomal PD-L1 and reduced surface PD-L1 levels, which led to defects in interaction with PD-1. Anti-PD-1 did not expand effector-like T cells within DRG2-depleted tumors and failed to improve the survival of DRG2-depleted tumor-bearing mice. Cohort analysis revealed that patients bearing melanoma with low DRG2 protein levels were resistant to anti-PD-1 therapy. These findings identify DRG2 as a key regulator of recycling of endosomal PD-L1 and response to anti-PD-1 therapy and provide insights into how to increase the correlation between PD-L1 expression and response to anti-PD-1 therapy.

Tm4sf19 deficiency inhibits osteoclast multinucleation and prevents bone loss.

BACKGROUND: Multinucleation is a hallmark of osteoclast formation and has a unique ability to resorb bone matrix. During osteoclast differentiation, the cytoskeleton reorganization results in the generation of actin belts and eventual bone resorption. Tetraspanins are involved in adhesion, migration and fusion in various cells. However, its function in osteoclast is still unclear. In this study, we identified Tm4sf19, a member of the tetraspanin family, as a regulator of osteoclast function. MATERIALS AND METHODS: We investigate the effect of Tm4sf19 deficiency on osteoclast differentiation using bone marrow-derived macrophages obtained from wild type (WT), Tm4sf19 knockout (KO) and Tm4sf19 LELΔ mice lacking the large extracellular loop (LEL). We analyzed bone mass of young and aged WT, KO and LELΔ mice by µCT analysis. The effects of Tm4sf19 LEL-Fc fusion protein were accessed in osteoclast differentiation and osteoporosis animal model. RESULTS: We found that deficiency of Tm4sf19 inhibited osteoclast function and LEL of Tm4sf19 was responsible for its function in osteoclasts in vitro. KO and LELΔ mice exhibited higher trabecular bone mass compared to WT mice. We found that Tm4sf19 interacts with integrin αvß3 through LEL, and that this binding is important for cytoskeletal rearrangements in osteoclast by regulating signaling downstream of integrin αvß3. Treatment with LEL-Fc fusion protein inhibited osteoclast function in vitro and administration of LEL-Fc prevented bone loss in an osteoporosis mouse model in vivo. CONCLUSION: We suggest that Tm4sf19 regulates osteoclast function and that LEL-Fc may be a promising drug to target bone destructive diseases caused by osteoclast hyper-differentiation.

Identification of Niche-Specific Gene Signatures between Malignant Tumor Microenvironments by Integrating Single Cell and Spatial Transcriptomics Data.

The tumor microenvironment significantly affects the transcriptomic states of tumor cells. Single-cell RNA sequencing (scRNA-seq) helps elucidate the transcriptomes of individual cancer cells and their neighboring cells. However, cell dissociation results in the loss of information on neighboring cells. To address this challenge and comprehensively assess the gene activity in tissue samples, it is imperative to integrate scRNA-seq with spatial transcriptomics. In our previous study on physically interacting cell sequencing (PIC-seq), we demonstrated that gene expression in single cells is affected by neighboring cell information. In the present study, we proposed a strategy to identify niche-specific gene signatures by harmonizing scRNA-seq and spatial transcriptomic data. This approach was applied to the paired or matched scRNA-seq and Visium platform data of five cancer types: breast cancer, gastrointestinal stromal tumor, liver hepatocellular carcinoma, uterine corpus endometrial carcinoma, and ovarian cancer. We observed distinct gene signatures specific to cellular niches and their neighboring counterparts. Intriguingly, these niche-specific genes display considerable dissimilarity to cell type markers and exhibit unique functional attributes independent of the cancer types. Collectively, these results demonstrate the potential of this integrative approach for identifying novel marker genes and their spatial relationships.

N-linked glycosylation is essential for anti-tumor activities of KIAA1324 in gastric cancer.

KIAA1324 is a transmembrane protein largely reported as a tumor suppressor and favorable prognosis marker in various cancers, including gastric cancer. In this study, we report the role of N-linked glycosylation in KIAA1324 as a functional post-translational modification (PTM). Loss of N-linked glycosylation eliminated the potential of KIAA1324 to suppress cancer cell proliferation and migration. Furthermore, we demonstrated that KIAA1324 undergoes fucosylation, a modification of the N-glycan mediated by fucosyltransferase, and inhibition of fucosylation also significantly suppressed KIAA1324-induced cell growth inhibition and apoptosis of gastric cancer cells. In addition, KIAA1324-mediated apoptosis and tumor regression were inhibited by the loss of N-linked glycosylation. RNA sequencing (RNAseq) analysis revealed that genes most relevant to the apoptosis and cell cycle arrest pathways were modulated by KIAA1324 with the N-linked glycosylation, and Gene Regulatory Network (GRN) analysis suggested novel targets of KIAA1324 for anti-tumor effects in the transcription level. The N-linked glycosylation blockade decreased protein stability through rapid proteasomal degradation. The non-glycosylated mutant also showed altered localization and lost apoptotic activity that inhibits the interaction between GRP78 and caspase 7. These data demonstrate that N-linked glycosylation of KIAA1324 is essential for the suppressive role of KIAA1324 protein in gastric cancer progression and indicates that KIAA1324 may have anti-tumor effects by targeting cancer-related genes with N-linked glycosylation. In conclusion, our study suggests the PTM of KIAA1324 including N-linked glycosylation and fucosylation is a necessary factor to consider for cancer prognosis and therapy improvement.