Deep learning tackles single-cell analysis-a survey of deep learning for scRNA-seq analysis.

Since its selection as the method of the year in 2013, single-cell technologies have become mature enough to provide answers to complex research questions. With the growth of single-cell profiling technologies, there has also been a significant increase in data collected from single-cell profilings, resulting in computational challenges to process these massive and complicated datasets. To address these challenges, deep learning (DL) is positioned as a competitive alternative for single-cell analyses besides the traditional machine learning approaches. Here, we survey a total of 25 DL algorithms and their applicability for a specific step in the single cell RNA-seq processing pipeline. Specifically, we establish a unified mathematical representation of variational autoencoder, autoencoder, generative adversarial network and supervised DL models, compare the training strategies and loss functions for these models, and relate the loss functions of these models to specific objectives of the data processing step. Such a presentation will allow readers to choose suitable algorithms for their particular objective at each step in the pipeline. We envision that this survey will serve as an important information portal for learning the application of DL for scRNA-seq analysis and inspire innovative uses of DL to address a broader range of new challenges in emerging multi-omics and spatial single-cell sequencing.

Wildfire smoke impacts lake ecosystems.

Wildfire activity is increasing globally. The resulting smoke plumes can travel hundreds to thousands of kilometers, reflecting or scattering sunlight and depositing particles within ecosystems. Several key physical, chemical, and biological processes in lakes are controlled by factors affected by smoke. The spatial and temporal scales of lake exposure to smoke are extensive and under-recognized. We introduce the concept of the lake smoke-day, or the number of days any given lake is exposed to smoke in any given fire season, and quantify the total lake smoke-day exposure in North America from 2019 to 2021. Because smoke can be transported at continental to intercontinental scales, even regions that may not typically experience direct burning of landscapes by wildfire are at risk of smoke exposure. We found that 99.3% of North America was covered by smoke, affecting a total of 1,333,687 lakes ≥10 ha. An incredible 98.9% of lakes experienced at least 10 smoke-days a year, with 89.6% of lakes receiving over 30 lake smoke-days, and lakes in some regions experiencing up to 4 months of cumulative smoke-days. Herein we review the mechanisms through which smoke and ash can affect lakes by altering the amount and spectral composition of incoming solar radiation and depositing carbon, nutrients, or toxic compounds that could alter chemical conditions and impact biota. We develop a conceptual framework that synthesizes known and theoretical impacts of smoke on lakes to guide future research. Finally, we identify emerging research priorities that can help us better understand how lakes will be affected by smoke as wildfire activity increases due to climate change and other anthropogenic activities.

Prediction and interpretation of cancer survival using graph convolution neural networks.

The survival rate of cancer has increased significantly during the past two decades for breast, prostate, testicular, and colon cancer, while the brain and pancreatic cancers have a much lower median survival rate that has not improved much over the last forty years. This has imposed the challenge of finding gene markers for early cancer detection and treatment strategies. Different methods including regression-based Cox-PH, artificial neural networks, and recently deep learning algorithms have been proposed to predict the survival rate for cancers. We established in this work a novel graph convolution neural network (GCNN) approach called Surv_GCNN to predict the survival rate for 13 different cancer types using the TCGA dataset. For each cancer type, 6 Surv_GCNN models with graphs generated by correlation analysis, GeneMania database, and correlation + GeneMania were trained with and without clinical data to predict the risk score (RS). The performance of the 6 Surv_GCNN models was compared with two other existing models, Cox-PH and Cox-nnet. The results showed that Cox-PH has the worst performance among 8 tested models across the 13 cancer types while Surv_GCNN models with clinical data reported the best overall performance, outperforming other competing models in 7 out of 13 cancer types including BLCA, BRCA, COAD, LUSC, SARC, STAD, and UCEC. A novel network-based interpretation of Surv_GCNN was also proposed to identify potential gene markers for breast cancer. The signatures learned by the nodes in the hidden layer of Surv_GCNN were identified and were linked to potential gene markers by network modularization. The identified gene markers for breast cancer have been compared to a total of 213 gene markers from three widely cited lists for breast cancer survival analysis. About 57% of gene markers obtained by Surv_GCNN with correlation + GeneMania graph either overlap or directly interact with the 213 genes, confirming the effectiveness of the identified markers by Surv_GCNN.

Drone phenotyping and machine learning enable discovery of loci regulating daily floral opening in lettuce.

Flower opening and closure are traits of reproductive importance in all angiosperms because they determine the success of self- and cross-pollination. The temporal nature of this phenotype rendered it a difficult target for genetic studies. Cultivated and wild lettuce, Lactuca spp., have composite inflorescences that open only once. An L. serriola×L. sativa F6 recombinant inbred line (RIL) population differed markedly for daily floral opening time. This population was used to map the genetic determinants of this trait; the floral opening time of 236 RILs was scored using time-course image series obtained by drone-based phenotyping on two occasions. Floral pixels were identified from the images using a support vector machine with an accuracy >99%. A Bayesian inference method was developed to extract the peak floral opening time for individual genotypes from the time-stamped image data. Two independent quantitative trait loci (QTLs; Daily Floral Opening 2.1 and qDFO8.1) explaining >30% of the phenotypic variation in floral opening time were discovered. Candidate genes with non-synonymous polymorphisms in coding sequences were identified within the QTLs. This study demonstrates the power of combining remote sensing, machine learning, Bayesian statistics, and genome-wide marker data for studying the genetics of recalcitrant phenotypes.

Deriving a Boolean dynamics to reveal macrophage activation with in vitro temporal cytokine expression profiles.

BACKGROUND: Macrophages show versatile functions in innate immunity, infectious diseases, and progression of cancers and cardiovascular diseases. These versatile functions of macrophages are conducted by different macrophage phenotypes classified as classically activated macrophages and alternatively activated macrophages due to different stimuli in the complex in vivo cytokine environment. Dissecting the regulation of macrophage activations will have a significant impact on disease progression and therapeutic strategy. Mathematical modeling of macrophage activation can improve the understanding of this biological process through quantitative analysis and provide guidance to facilitate future experimental design. However, few results have been reported for a complete model of macrophage activation patterns. RESULTS: We globally searched and reviewed literature for macrophage activation from PubMed databases and screened the published experimental results. Temporal in vitro macrophage cytokine expression profiles from published results were selected to establish Boolean network models for macrophage activation patterns in response to three different stimuli. A combination of modeling methods including clustering, binarization, linear programming (LP), Boolean function determination, and semi-tensor product was applied to establish Boolean networks to quantify three macrophage activation patterns. The structure of the networks was confirmed based on protein-protein-interaction databases, pathway databases, and published experimental results. Computational predictions of the network evolution were compared against real experimental results to validate the effectiveness of the Boolean network models. CONCLUSION: Three macrophage activation core evolution maps were established based on the Boolean networks using Matlab. Cytokine signatures of macrophage activation patterns were identified, providing a possible determination of macrophage activations using extracellular cytokine measurements.

Transformative Impact of Proteomics on Cardiovascular Health and Disease: A Scientific Statement From the American Heart Association.

The year 2014 marked the 20th anniversary of the coining of the term proteomics. The purpose of this scientific statement is to summarize advances over this period that have catalyzed our capacity to address the experimental, translational, and clinical implications of proteomics as applied to cardiovascular health and disease and to evaluate the current status of the field. Key successes that have energized the field are delineated; opportunities for proteomics to drive basic science research, facilitate clinical translation, and establish diagnostic and therapeutic healthcare algorithms are discussed; and challenges that remain to be solved before proteomic technologies can be readily translated from scientific discoveries to meaningful advances in cardiovascular care are addressed. Proteomics is the result of disruptive technologies, namely, mass spectrometry and database searching, which drove protein analysis from 1 protein at a time to protein mixture analyses that enable large-scale analysis of proteins and facilitate paradigm shifts in biological concepts that address important clinical questions. Over the past 20 years, the field of proteomics has matured, yet it is still developing rapidly. The scope of this statement will extend beyond the reaches of a typical review article and offer guidance on the use of next-generation proteomics for future scientific discovery in the basic research laboratory and clinical settings.

Detection of high variability in gene expression from single-cell RNA-seq profiling.

BACKGROUND: The advancement of the next-generation sequencing technology enables mapping gene expression at the single-cell level, capable of tracking cell heterogeneity and determination of cell subpopulations using single-cell RNA sequencing (scRNA-seq). Unlike the objectives of conventional RNA-seq where differential expression analysis is the integral component, the most important goal of scRNA-seq is to identify highly variable genes across a population of cells, to account for the discrete nature of single-cell gene expression and uniqueness of sequencing library preparation protocol for single-cell sequencing. However, there is lack of generic expression variation model for different scRNA-seq data sets. Hence, the objective of this study is to develop a gene expression variation model (GEVM), utilizing the relationship between coefficient of variation (CV) and average expression level to address the over-dispersion of single-cell data, and its corresponding statistical significance to quantify the variably expressed genes (VEGs). RESULTS: We have built a simulation framework that generated scRNA-seq data with different number of cells, model parameters, and variation levels. We implemented our GEVM and demonstrated the robustness by using a set of simulated scRNA-seq data under different conditions. We evaluated the regression robustness using root-mean-square error (RMSE) and assessed the parameter estimation process by varying initial model parameters that deviated from homogeneous cell population. We also applied the GEVM on real scRNA-seq data to test the performance under distinct cases. CONCLUSIONS: In this paper, we proposed a gene expression variation model that can be used to determine significant variably expressed genes. Applying the model to the simulated single-cell data, we observed robust parameter estimation under different conditions with minimal root mean square errors. We also examined the model on two distinct scRNA-seq data sets using different single-cell protocols and determined the VEGs. Obtaining VEGs allowed us to observe possible subpopulations, providing further evidences of cell heterogeneity. With the GEVM, we can easily find out significant variably expressed genes in different scRNA-seq data sets.

Translating Koch's postulates to identify matrix metalloproteinase roles in postmyocardial infarction remodeling: cardiac metalloproteinase actions (CarMA) postulates.

The first matrix metalloproteinase (MMP) was described in 1962; and since the 1990s, cardiovascular research has focused on understanding how MMPs regulate many aspects of cardiovascular pathology from atherosclerosis formation to myocardial infarction and stroke. Although much information has been gleaned by these past reports, to a large degree MMP cardiovascular biology remains observational, with few studies homing in on cause and effect relationships. Koch's postulates were first developed in the 19th century as a way to establish microorganism function and were modified in the 20th century to include methods to establish molecular causality. In this review, we outline the concept for establishing a similar approach to determine causality in terms of MMP functions. We use left ventricular remodeling postmyocardial infarction as an example, but this approach will have broad applicability across both the cardiovascular and the MMP fields.

Systems analysis of gene ontology and biological pathways involved in post-myocardial infarction responses.

BACKGROUND: Pathway analysis has been widely used to gain insight into essential mechanisms of the response to myocardial infarction (MI). Currently, there exist multiple pathway databases that organize molecular datasets and manually curate pathway maps for biological interpretation at varying forms of organization. However, inconsistencies among different databases in pathway descriptions, frequently due to conflicting results in the literature, can generate incorrect interpretations. Furthermore, although pathway analysis software provides detailed images of interactions among molecules, it does not exhibit how pathways interact with one another or with other biological processes under specific conditions. METHODS: We propose a novel method to standardize descriptions of enriched pathways for a set of genes/proteins using Gene Ontology terms. We used this method to examine the relationships among pathways and biological processes for a set of condition-specific genes/proteins, represented as a functional biological pathway-process network. We applied this algorithm to a set of 613 MI-specific proteins we previously identified. RESULTS: A total of 96 pathways from Biocarta, KEGG, and Reactome, and 448 Gene Ontology Biological Processes were enriched with these 613 proteins. The pathways were represented as Boolean functions of biological processes, delivering an interactive scheme to organize enriched information with an emphasis on involvement of biological processes in pathways. We extracted a network focusing on MI to demonstrate that tyrosine phosphorylation of Signal Transducer and Activator of Transcription (STAT) protein, positive regulation of collagen metabolic process, coagulation, and positive/negative regulation of blood coagulation have immediate impacts on the MI response. CONCLUSIONS: Our method organized biological processes and pathways in an unbiased approach to provide an intuitive way to identify biological properties of pathways under specific conditions. Pathways from different databases have similar descriptions yet diverse biological processes, indicating variation in their ability to share similar functional characteristics. The coverages of pathways can be expanded with the incorporation of more biological processes, predicting involvement of protein members in pathways. Further, detailed analyses of the functional biological pathway-process network will allow researchers and scientists to explore critical routes in biological systems in the progression of disease.

Plasma Glycoproteomics Reveals Sepsis Outcomes Linked to Distinct Proteins in Common Pathways.

OBJECTIVE: Sepsis remains a predominant cause of mortality in the ICU, yet strategies to increase survival have proved largely unsuccessful. This study aimed to identify proteins linked to sepsis outcomes using a glycoproteomic approach to target extracellular proteins that trigger downstream pathways and direct patient outcomes. DESIGN: Plasma was obtained from the Lactate Assessment in the Treatment of Early Sepsis cohort. N-linked plasma glycopeptides were quantified by solid-phase extraction coupled with mass spectrometry. Glycopeptides were assigned to proteins using RefSeq (National Center of Biotechnology Information, Bethesda, MD) and visualized in a heat map. Protein differences were validated by immunoblotting, and proteins were mapped for biological processes using Database for Annotation, Visualization and Integrated Discovery (National Institute of Allergy and Infectious Diseases, National Institutes of Health; Bethesda, MD) and for functional pathways using Kyoto Encyclopedia of Genes and Genomes (Kanehisa Laboratories, Kyoto, Japan) databases. SETTING: Hospitalized care. PATIENTS: Patients admitted to the emergency department were enrolled in the study when the diagnosis of sepsis was made, within 6 hours of presentation. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 501 glycopeptides corresponding to 234 proteins were identified. Of these, 66 glycopeptides were unique to the survivor group and corresponded to 54 proteins, 60 were unique to the nonsurvivor group and corresponded to 43 proteins, and 375 were common responses between groups and corresponded to 137 proteins. Immunoblotting showed that nonsurvivors had increased total kininogen; decreased total cathepsin-L1, vascular cell adhesion molecule, periostin, and neutrophil gelatinase-associated lipocalin; and a two-fold decrease in glycosylated clusterin (all p < 0.05). Kyoto Encyclopedia of Genes and Genomes analysis identified six enriched pathways. Interestingly, survivors relied on the extrinsic pathway of the complement and coagulation cascade, whereas nonsurvivors relied on the intrinsic pathway. CONCLUSION: This study identifies proteins linked to patient outcomes and provides insight into unexplored mechanisms that can be investigated for the identification of novel therapeutic targets.

Matrix metalloproteinase-28 deletion exacerbates cardiac dysfunction and rupture after myocardial infarction in mice by inhibiting M2 macrophage activation.

RATIONALE: Matrix metalloproteinase (MMP)-28 regulates the inflammatory and extracellular matrix responses in cardiac aging, but the roles of MMP-28 after myocardial infarction (MI) have not been explored. OBJECTIVE: To determine the impact of MMP-28 deletion on post-MI remodeling of the left ventricle (LV). METHODS AND RESULTS: Adult C57BL/6J wild-type (n=76) and MMP null (MMP-28((-/-)), n=86) mice of both sexes were subjected to permanent coronary artery ligation to create MI. MMP-28 expression decreased post-MI, and its cell source shifted from myocytes to macrophages. MMP-28 deletion increased day 7 mortality because of increased cardiac rupture post-MI. MMP-28(-/-) mice exhibited larger LV volumes, worse LV dysfunction, a worse LV remodeling index, and increased lung edema. Plasma MMP-9 levels were unchanged in the MMP-28((-/-)) mice but increased in wild-type mice at day 7 post-MI. The mRNA levels of inflammatory and extracellular matrix proteins were attenuated in the infarct regions of MMP-28(-/-) mice, indicating reduced inflammatory and extracellular matrix responses. M2 macrophage activation was impaired when MMP-28 was absent. MMP-28 deletion also led to decreased collagen deposition and fewer myofibroblasts. Collagen cross-linking was impaired as a result of decreased expression and activation of lysyl oxidase in the infarcts of MMP-28(-/-) mice. The LV tensile strength at day 3 post-MI, however, was similar between the 2 genotypes. CONCLUSIONS: MMP-28 deletion aggravated MI-induced LV dysfunction and rupture as a result of defective inflammatory response and scar formation by suppressing M2 macrophage activation.

Integrative computational and experimental approaches to establish a post-myocardial infarction knowledge map.

Vast research efforts have been devoted to providing clinical diagnostic markers of myocardial infarction (MI), leading to over one million abstracts associated with "MI" and "Cardiovascular Diseases" in PubMed. Accumulation of the research results imposed a challenge to integrate and interpret these results. To address this problem and better understand how the left ventricle (LV) remodels post-MI at both the molecular and cellular levels, we propose here an integrative framework that couples computational methods and experimental data. We selected an initial set of MI-related proteins from published human studies and constructed an MI-specific protein-protein-interaction network (MIPIN). Structural and functional analysis of the MIPIN showed that the post-MI LV exhibited increased representation of proteins involved in transcriptional activity, inflammatory response, and extracellular matrix (ECM) remodeling. Known plasma or serum expression changes of the MIPIN proteins in patients with MI were acquired by data mining of the PubMed and UniProt knowledgebase, and served as a training set to predict unlabeled MIPIN protein changes post-MI. The predictions were validated with published results in PubMed, suggesting prognosticative capability of the MIPIN. Further, we established the first knowledge map related to the post-MI response, providing a major step towards enhancing our understanding of molecular interactions specific to MI and linking the molecular interaction, cellular responses, and biological processes to quantify LV remodeling.

P. gingivalis lipopolysaccharide intensifies inflammation post-myocardial infarction through matrix metalloproteinase-9.

Periodontal disease (PD) strongly correlates with increased mortality post-myocardial infarction (MI); however, the underlying mechanisms are unknown. Matrix metalloproteinase (MMP)-9 levels directly correlate with dysfunction and remodeling of the left ventricle (LV) post-MI. Post-MI, MMP-9 is produced by leukocytes and modulates inflammation. We have shown that exposure to Porphyromonas gingivalis lipopolysaccharide (PgLPS), an immunomodulatory molecule identified in PD patients, increases LV MMP-9 levels in mice and leads to cardiac inflammation and dysfunction. The aim of the study was to determine if circulating PgLPS exacerbates the LV inflammatory response post-MI through MMP-9 dependent mechanisms. We exposed wild type C57BL/6J and MMP-9(-/-) mice to PgLPS (ATCC 33277) for a period of 28 days before performing MI, and continued to deliver PgLPS for up to 7 days post-MI. We found systemic levels of PgLPS 1) increased MMP-9 levels in both plasma and infarcted LV resulting in reduced wall thickness and increased incidence of LV rupture post-MI and 2) increased systemic and local macrophage chemotaxis leading to accelerated M1 macrophage infiltration post-MI and decreased LV function. MMP-9 deletion played a protective role by attenuating the inflammation induced by systemic delivery of PgLPS. In conclusion, MMP-9 deletion has a cardioprotective role against PgLPS exposure, by attenuating macrophage mediated inflammation.

Aliskiren and valsartan mediate left ventricular remodeling post-myocardial infarction in mice through MMP-9 effects.

We evaluated whether aliskiren, valsartan, or a combination of both was protective following myocardial infarction (MI) through effects on matrix metalloproteinase (MMP)-9. C57BL/6J wild type (WT, n=94) and MMP-9 null (null, n=85) mice were divided into 4 groups at 3h post-MI: saline (S), aliskiren (A; 50mg/kg/day), valsartan (V; 40mg/kg/day), or A+V and compared to no MI controls at 28days post-MI. All groups had similar infarct areas, and survival rates were higher in the null mice. The treatments influenced systolic function and hypertrophy index, as well as extracellular matrix (ECM) and inflammatory genes in the remote region, indicating that primary effects were on the viable myocardium. Saline treated WT mice showed increased end systolic and diastolic volumes and hypertrophy index, along with reduced ejection fraction. MMP-9 deletion improved LV function post-MI. Aliskiren attenuated the increase in end systolic volume and hypertrophy index, while valsartan improved end diastolic volumes and aliskiren+valsartan improved the hypertrophy index only when MMP-9 was absent. Extracellular matrix and inflammatory gene expression showed distinct patterns among the treatment groups, indicating a divergence in mechanisms of remodeling. This study shows that MMP-9 regulates aliskiren and valsartan effects in mice. These results in mice provide mechanistic insight to help translate these findings to post-MI patients.

Cardiac aging is initiated by matrix metalloproteinase-9-mediated endothelial dysfunction.

Aging is linked to increased matrix metalloproteinase-9 (MMP-9) expression and extracellular matrix turnover, as well as a decline in function of the left ventricle (LV). Previously, we demonstrated that C57BL/6J wild-type (WT) mice > 18 mo of age show impaired diastolic function, which was attenuated by MMP-9 deletion. To evaluate mechanisms that initiate the development of cardiac dysfunction, we compared the LVs of 6-9- and 15-18-mo-old WT and MMP-9 null (Null) mice. All groups showed similar LV function by echocardiography, indicating that dysfunction had not yet developed in the older group. Myocyte nuclei numbers and cross-sectional areas increased in both WT and Null 15-18-mo mice compared with young controls, indicating myocyte hypertrophy. Myocyte hypertrophy leads to an increased oxygen demand, and both WT and Null 15-18-mo mice showed an increase in angiogenic signaling. Plasma proteomic profiling and LV analysis revealed a threefold increase in von Willebrand factor and fivefold increase in vascular endothelial growth factor in WT 15-18-mo mice, which were further elevated in Null mice. In contrast to the upregulation of angiogenic stimulating factors, actual LV vessel numbers increased only in the 15-18-mo Null LV. The 15-18-mo WT showed amplified expression of inflammatory genes related to angiogenesis, including C-C chemokine receptor (CCR)7, CCR10, interleukin (IL)-1f8, IL-13, and IL-20 (all, P < 0.05), and these increases were blunted by MMP-9 deletion (all, P < 0.05). To measure vascular permeability as an index of endothelial function, we injected mice with FITC-labeled dextran. The 15-18-mo WT LV showed increased vascular permeability compared with young WT controls and 15-18-mo Null mice. Combined, our findings revealed that MMP-9 deletion improves angiogenesis, attenuates inflammation, and prevents vascular leakiness in the setting of cardiac aging.

Vegetation controls on northern high latitude snow-albedo feedback: observations and CMIP5 model simulations.

The snow-masking effect of vegetation exerts strong control on albedo in northern high latitude ecosystems. Large-scale changes in the distribution and stature of vegetation in this region will thus have important feedbacks to climate. The snow-albedo feedback is controlled largely by the contrast between snow-covered and snow-free albedo (Δα), which influences predictions of future warming in coupled climate models, despite being poorly constrained at seasonal and century time scales. Here, we compare satellite observations and coupled climate model representations of albedo and tree cover for the boreal and Arctic region. Our analyses reveal consistent declines in albedo with increasing tree cover, occurring south of latitudinal tree line, that are poorly represented in coupled climate models. Observed relationships between albedo and tree cover differ substantially between snow-covered and snow-free periods, and among plant functional type. Tree cover in models varies widely but surprisingly does not correlate well with model albedo. Furthermore, our results demonstrate a relationship between tree cover and snow-albedo feedback that may be used to accurately constrain high latitude albedo feedbacks in coupled climate models under current and future vegetation distributions.

Reduced BDNF attenuates inflammation and angiogenesis to improve survival and cardiac function following myocardial infarction in mice.

Brain-derived neurotrophic factor (BDNF) increases in failing hearts, but BDNF roles in cardiac remodeling following myocardial infarction (MI) are unclear. Male BDNF(+/+) [wild-type (WT)] and BDNF(+/-) heterozygous (HET) mice at 6-9 mo of age were subjected to MI and evaluated at days 1, 3, 5, 7, or 28 post-MI. At day 28 post-MI, 76% of HET versus 40% of WT survived, whereas fractional shortening improved and neovascularization levels were reduced in the HET (all, P < 0.05). At day 1, post-MI, matrix metalloproteinase-9, and myeloperoxidase (MPO) increased in WT, but not in HET. Concomitantly, monocyte chemotactic protein-1 and -5 levels increased and vascular endothelial growth factor (VEGF)-A decreased in HET. Neutrophil infiltration peaked at days 1-3 in WT mice, and this increase was blunted in HET. To determine if MPO administration could rescue the HET phenotype, MPO was injected at 3 h post-MI. MPO restored VEGF-A levels without altering matrix metalloproteinase-9 or neutrophil content. In conclusion, reduced BDNF levels modulated the early inflammatory and neovascularization responses, leading to improved survival and reduced cardiac remodeling at day 28 post-MI. Thus reduced BDNF attenuates early inflammation following MI by modulating MPO and angiogenic response through VEGF-A.

Estimation of tomato water status with photochemical reflectance index and machine learning: Assessment from proximal sensors and UAV imagery.

Tracking plant water status is a critical step towards the adaptive precision irrigation management of processing tomatoes, one of the most important specialty crops in California. The photochemical reflectance index (PRI) from proximal sensors and the high-resolution unmanned aerial vehicle (UAV) imagery provide an opportunity to monitor the crop water status efficiently. Based on data from an experimental tomato field with intensive aerial and plant-based measurements, we developed random forest machine learning regression models to estimate tomato stem water potential (ψ stem), (using observations from proximal sensors and 12-band UAV imagery, respectively, along with weather data. The proximal sensor-based model estimation agreed well with the plant ψ stem with R 2 of 0.74 and mean absolute error (MAE) of 0.63 bars. The model included PRI, normalized difference vegetation index, vapor pressure deficit, and air temperature and tracked well with the seasonal dynamics of ψ stem across different plots. A separate model, built with multiple vegetation indices (VIs) from UAV imagery and weather variables, had an R 2 of 0.81 and MAE of 0.67 bars. The plant-level ψ stem maps generated from UAV imagery closely represented the water status differences of plots under different irrigation treatments and also tracked well the temporal change among flights. PRI was found to be the most important VI in both the proximal sensor- and the UAV-based models, providing critical information on tomato plant water status. This study demonstrated that machine learning models can accurately estimate the water status by integrating PRI, other VIs, and weather data, and thus facilitate data-driven irrigation management for processing tomatoes.

Tree-level almond yield estimation from high resolution aerial imagery with convolutional neural network.

Introduction: Estimating and understanding the yield variability within an individual field is critical for precision agriculture resource management of high value tree crops. Recent advancements in sensor technologies and machine learning make it possible to monitor orchards at very high spatial resolution and estimate yield at individual tree level. Methods: This study evaluates the potential of utilizing deep learning methods to predict tree-level almond yield with multi-spectral imagery. We focused on an almond orchard with the 'Independence' cultivar in California, where individual tree harvesting and yield monitoring was conducted for ~2,000 trees and summer aerial imagery at 30cm was acquired for four spectral bands in 2021. We developed a Convolutional Neural Network (CNN) model with a spatial attention module to take the multi-spectral reflectance imagery directly for almond fresh weight estimation at the tree level. Results: The deep learning model was shown to predict the tree level yield very well, with a R2 of 0.96 (±0.002) and Normalized Root Mean Square Error (NRMSE) of 6.6% (±0.2%), based on 5-fold cross validation. The CNN estimation captured well the patterns of yield variation between orchard rows, along the transects, and from tree to tree, when compared to the harvest data. The reflectance at the red edge band was found to play the most important role in the CNN yield estimation. Discussion: This study demonstrates the significant improvement of deep learning over traditional linear regression and machine learning methods for accurate and robust tree level yield estimation, highlighting the potential for data-driven site-specific resource management to ensure agriculture sustainability.

Transcriptional profiles of non-neuronal and immune cells in mouse trigeminal ganglia.

Non-neuronal cells constitute 90%-95% of sensory ganglia. These cells, especially glial and immune cells, play critical roles in the modulation of sensory neurons. This study aimed to identify, profile, and summarize the types of trigeminal ganglion (TG) non-neuronal cells in naïve male mice using published and our own data generated by single-cell RNA sequencing, flow cytometry, and immunohistochemistry. TG has five types of non-neuronal cells, namely, glial, fibroblasts, smooth muscle, endothelial, and immune cells. There is an agreement among publications for glial, fibroblasts, smooth muscle, and endothelial cells. Based on gene profiles, glial cells were classified as myelinated and non-myelinated Schwann cells and satellite glial cells. Mpz has dominant expression in Schwann cells, and Fabp7 is specific for SCG. Two types of Col1a2+ fibroblasts located throughout TG were distinguished. TG smooth muscle and endothelial cells in the blood vessels were detected using well-defined markers. Our study reported three types of macrophages (Mph) and four types of neutrophils (Neu) in TG. Mph were located in the neuronal bodies and nerve fibers and were sub-grouped by unique transcriptomic profiles with Ccr2, Cx3cr1, and Iba1 as markers. A comparison of databases showed that type 1 Mph is similar to choroid plexus-low (CPlo) border-associated Mph (BAMs). Type 2 Mph has the highest prediction score with CPhi BAMs, while type 3 Mph is distinct. S100a8+ Neu were located in the dura surrounding TG and were sub-grouped by clustering and expressions of Csf3r, Ly6G, Ngp, Elane, and Mpo. Integrative analysis of published datasets indicated that Neu-1, Neu-2, and Neu-3 are similar to the brain Neu-1 group, while Neu-4 has a resemblance to the monocyte-derived cells. Overall, the generated and summarized datasets on non-neuronal TG cells showed a unique composition of myeloid cell types in TG and could provide essential and fundamental information for studies on cell plasticity, interactomic networks between neurons and non-neuronal cells, and function during a variety of pain conditions in the head and neck regions.