Control Dataset Research Articles

Bi-directional effects between inflammatory molecules and intracranial aneurysm.

Although inflammation is closely associated with the pathogenesis of intracranial aneurysm (IA), detailed causal associations remain unclear. This study aimed to investigate the causality between circulating inflammatory molecules (IMs) and IA. The bi-directional Mendelian randomization (MR) analysis was conducted using two genome-wide association studies (GWAS) for inflammatory molecules (IMs) from Finnish and Icelandic populations, as well as GWAS datasets of IA cases and controls of European descent. Colocalization analysis was performed to validate MR associations. Subsequently, Venn analysis was conducted to identify the overlapped causalities. Integrating the findings from two MR models, RANTES was suggestively associated with IA (Finnish model: inverse variance-weighted odds ratio [ORIVW] (95% confidence interval [95% CI]), 0.86 (0.74-0.99); Icelandic model: 0.80 (0.68-0.94)) and aneurysmal subarachnoid hemorrhage (aSAH) (ORIVW (95% CI): 0.81 (0.68-0.95) and 0.80 (0.66-0.97) in Finnish and Icelandic models). IA and its subtypes were not associated with any of candidate IMs. However, colocalization analysis failed to identify significant evidence of shared genetic instruments between the exposures and outcomes, except for the MCP3-aSAH pair in the Icelandic model. No significant causality was identified between IMs and IA or their subtypes. RANTES is potentially associated with IA and aSAH. Further investigation is warranted to explore the role of IMs in IA development.

Networks extracted from nonlinear fMRI connectivity exhibit unique spatial variation and enhanced sensitivity to differences between individuals with schizophrenia and controls

Schizophrenia is a chronic brain disorder associated with widespread alterations in functional brain connectivity. Although data-driven approaches such as independent component analysis are often used to study how schizophrenia impacts linearly connected networks, alterations within the underlying nonlinear functional connectivity structure remain largely unknown. Here we report the analysis of networks from explicitly nonlinear functional magnetic resonance imaging connectivity in a case–control dataset. We found systematic spatial variation, with higher nonlinear weight within core regions, suggesting that linear analyses underestimate functional connectivity within network centers. We also found that a unique nonlinear network incorporating default-mode, cingulo-opercular and central executive regions exhibits hypoconnectivity in schizophrenia, indicating that typically hidden connectivity patterns may reflect inefficient network integration in psychosis. Moreover, nonlinear networks including those previously implicated in auditory, linguistic and self-referential cognition exhibit heightened statistical sensitivity to schizophrenia diagnosis, collectively underscoring the potential of our methodology to resolve complex brain phenomena and transform clinical connectivity analysis.

QTL mapping and transcriptome analysis of seed germination under PEG-induced water stress in Lactuca spp.

The impact of limited water availability on lettuce growth has been well documented. However, the mechanisms by which lettuce controls seed germination under water stress remain unknown. Germination percentage was evaluated in the cv. Salinas (Lactuca sativa) (L. sativa) × US96UC23 (Lactuca serriola) (L. serriola) recombinant inbred line (RIL) population and USDA germplasm collection using 10% polyethylene glycol (PEG). About 50% of both populations displayed less than 90% germination. The average broad-sense heritability (H2) for germination percentage was 0.81 across both populations. Two quantitative trait loci (QTL) for germination percentage were identified on chromosomes 4 and 8 in the RIL population. The RNA-Seq and network analyses of wild lettuce, US96UC23, were performed using the control (distilled water, dH2O) and treatment (10% PEG) datasets. The number of differentially expressed genes (DEGs) was 4,095. The top 20 gene ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were assessed by enrichment analysis. The consensus network analysis captured 44 modules. Gene networks were constructed for the top 20 hub genes in 10 significant modules from each dataset. This study comprehensively explains QTL, GO terms, KEGG pathways, and gene networks associated with lettuce seed germination under osmotic stress.

How to evaluate exertional breathlessness using normative reference equations in research.

Breathlessness is a common, distressing and limiting symptom in people with advanced disease, but is challenging to assess as the symptom intensity depends on the level of exertion (symptom stimulus) during the assessment. This review outlines how to use recently developed normative reference equations to evaluate breathlessness responses, accounting for level of exertion, for valid assessment in symptom research. Published normative reference equations are freely available to predict the breathlessness intensity response (on a 0-10 Borg scale) among healthy people after a 6-minute walking test (6MWT) or an incremental cycle cardiopulmonary exercise test (iCPET). The predicted normal values account for individual characteristics (including age, sex, height, and body mass) and level of exertion (walk distance for 6MWT; power output, oxygen uptake, or minute ventilation at any point during the iCPET). The equations can be used to (1) construct a matched healthy control dataset for a study; (2) determine how abnormal an individual's exertional breathlessness is compared with healthy controls; (3) identify abnormal exertional breathlessness (rating>upper limit of normal); and (4) validly compare exertional breathlessness levels across individuals and groups. Methods for standardized and valid assessment of exertional breathlessness have emerged for improved symptoms research.

Landscape of intrinsically disordered proteins in mental disorder diseases

Disrupted genes linked to mental disorders sometimes exhibit characteristics of Intrinsically Disordered Proteins (IDPs). However, few studies have comprehensively explored the functional associations between protein disorder properties and different psychiatric disorders. In this study, we collected disrupted proteins for seven mental diseases (MDD, SCZ, BP, ID, AD, ADHD, ASD) and a control dataset from normal brains. After calculating the disorder scores for each protein, we thoroughly compared the proportions and functions of IDPs between differentially expressed proteins in each disease and healthy controls.Our findings revealed that disrupted proteins, particularly in ASD and ADHD, contain more IDPs than controls from normal brains. Distinct patterns in disorder properties were observed among different mental disorders. Functional enrichment analysis indicated that IDPs in mental disorders were associated with neurodevelopment, synaptic signaling, and gene expression regulatory pathways. In addition, we analyzed the proportion and function of liquid-phase-separated proteins (LLPS) in psychiatric disorders, finding that LLPS proteins are mainly enriched in pathways related to neurodevelopment and inter-synaptic signaling. Furthermore, to validate our findings, we conducted an analysis of differentially expressed genes in an ASD cohort, revealing that the encoded proteins also exhibit a higher proportion of IDPs. Notably, these IDPs were particularly enriched in pathways related to neurodevelopment, including head development, a process known to be disrupted in ASD.Our study sheds light on the crucial role of IDPs in psychiatric disorders, enhancing our understanding of their molecular mechanisms.

Identification of positions in human aldolase a that are neutral for apparent KM

According to evolutionary theory, many naturally-occurring amino acid substitutions are expected to be neutral or near-neutral, with little effect on protein structure or function. Accordingly, most changes observed in human exomes are also expected to be neutral. As such, accurate algorithms for identifying medically-relevant changes must discriminate rare, non-neutral substitutions against a background of neutral substitutions. However, due to historical biases in biochemical experiments, the data available to train and validate prediction algorithms mostly contains non-neutral substitutions, with few examples of neutral substitutions. Thus, available training sets have the opposite composition of the desired test sets. Towards improving a dataset of these critical negative controls, we have concentrated on identifying neutral positions – those positions for which most of the possible 19 amino acid substitutions have little effect on protein structure or function. Here, we used a strategy based on multiple sequence alignments to identify putative neutral positions in human aldolase A, followed by biochemical assays for 147 aldolase substitutions. Results showed that most variants had little effect on either the apparent Michaelis constant for substrate fructose-1,6-bisphosphate or its apparent cooperativity. Thus, these data are useful for training and validating prediction algorithms. In addition, we created a database of these and other biochemically characterized aldolase variants along with aldolase sequences and characteristics derived from sequence and structure analyses. This database is publicly available at https://github.com/liskinsk/Aldolase-variant-and-sequence-database.

Imputation methods for mixed datasets in bioarchaeology

Missing data is a prevalent problem in bioarchaeological research and imputation could provide a promising solution. This work simulated missingness on a control dataset (481 samples × 41 variables) in order to explore imputation methods for mixed data (qualitative and quantitative data). The tested methods included Random Forest (RF), PCA/MCA, factorial analysis for mixed data (FAMD), hotdeck, predictive mean matching (PMM), random samples from observed values (RSOV), and a multi-method (MM) approach for the three missingness mechanisms (MCAR, MAR, and MNAR) at levels of 5%, 10%, 20%, 30%, and 40% missingness. This study also compared single imputation with an adapted multiple imputation method derived from the R package “mice”. The results showed that the adapted multiple imputation technique always outperformed single imputation for the same method. The best performing methods were most often RF and MM, and other commonly successful methods were PCA/MCA and PMM multiple imputation. Across all criteria, the amount of missingness was the most important parameter for imputation accuracy. While this study found that some imputation methods performed better than others for the control dataset, each imputation method has advantages and disadvantages. Imputation remains a promising solution for datasets containing missingness; however when making a decision it is essential to consider dataset structure and research goals.

Decoding sound content in the early visual cortex of aphantasic participants

Listening to natural auditory scenes leads to distinct neuronal activity patterns in the early visual cortex (EVC) of blindfolded sighted and congenitally blind participants.1,2 This pattern of sound decoding is organized by eccentricity, with the accuracy of auditory information increasing from foveal to far peripheral retinotopic regions in the EVC (V1, V2, and V3). This functional organization by eccentricity is predicted by primate anatomical connectivity,3,4 where cortical feedback projections from auditory and other non-visual areas preferentially target the periphery of early visual areas. In congenitally blind participants, top-down feedback projections to the visual cortex proliferate,5 which might account for even higher sound-decoding accuracy in the EVC compared with blindfolded sighted participants.2 In contrast, studies in participants with aphantasia suggest an impairment of feedback projections to early visual areas, leading to a loss of visual imagery experience.6,7 This raises the question of whether impaired visual feedback pathways in aphantasia also reduce the transmission of auditory information to early visual areas. We presented auditory scenes to 23 blindfolded aphantasic participants. We found overall decreased sound decoding in early visual areas compared to blindfolded sighted ("control") and blind participants. We further explored this difference by modeling eccentricity effects across the blindfolded control, blind, and aphantasia datasets, and with a whole-brain searchlight analysis. Our findings suggest that the feedback of auditory content to the EVC is reduced in aphantasic participants. Reduced top-down projections might lead to both less sound decoding and reduced subjective experience of visual imagery.

High-coverage nanopore sequencing of samples from the 1000 Genomes Project to build a comprehensive catalog of human genetic variation.

Fewer than half of individuals with a suspected Mendelian or monogenic condition receive a precise molecular diagnosis after comprehensive clinical genetic testing. Improvements in data quality and costs have heightened interest in using long-read sequencing (LRS) to streamline clinical genomic testing, but the absence of control datasets for variant filtering and prioritization has made tertiary analysis of LRS data challenging. To address this, the 1000 Genomes Project ONT Sequencing Consortium aims to generate LRS data from at least 800 of the 1000 Genomes Project samples. Our goal is to use LRS to identify a broader spectrum of variation so we may improve our understanding of normal patterns of human variation. Here, we present data from analysis of the first 100 samples, representing all 5 superpopulations and 19 subpopulations. These samples, sequenced to an average depth of coverage of 37x and sequence read N50 of 54 kbp, have high concordance with previous studies for identifying single nucleotide and indel variants outside of homopolymer regions. Using multiple structural variant (SV) callers, we identify an average of 24,543 high-confidence SVs per genome, including shared and private SVs likely to disrupt gene function as well as pathogenic expansions within disease-associated repeats that were not detected using short reads. Evaluation of methylation signatures revealed expected patterns at known imprinted loci, samples with skewed X-inactivation patterns, and novel differentially methylated regions. All raw sequencing data, processed data, and summary statistics are publicly available, providing a valuable resource for the clinical genetics community to discover pathogenic SVs.

Identification of necroptosis genes and characterization of immune infiltration in non-alcoholic steatohepatitis

BackgroundThe most common progressive form of non-alcoholic fatty liver disease (NAFLD) is non-alcoholic steatohepatitis (NASH), which is characterized by the development of cirrhosis, and requires liver transplantation. We screened for the differentially expressed necroptosis-related genes in NASH in this study, and analyzed immune infiltration through microarray and bioinformatics analysis to identify potential biomarkers, and explore the molecular mechanisms involved in NASH.MethodsThe GSE24807 microarray dataset of NASH patients and healthy controls was downloaded, and we identified the differentially expressed genes (DEGs). Necroptosis-related differential genes (NRDEGs) were extracted from these DEGs, and functionally annotated by enrichment analyses. The core genes were obtained by constructing gene co-expression networks using weighted gene co-expression network analysis (WGCNA). Finally, the transcription factor (TF) regulatory network and the mRNA-miRNA network were constructed, and the infiltrating immune cell populations were analyzed with CIBERSORT.ResultsWe identified six necroptosis-related genes (CASP1, GLUL, PYCARD, IL33, SHARPIN, and IRF9), and they are potential diagnostic biomarkers for NASH. In particular, PYCARD is a potential biomarker for NAFLD progression. Analyses of immune infiltration showed that M2 macrophages, γδ T cells, and T follicular helper cells were associated with the immune microenvironment of NASH, which is possibly regulated by CASP1, IL33, and IRF9.ConclusionsWe identified six necroptosis-related genes in NASH, which are also potential diagnostic biomarkers. Our study provides new insights into the molecular mechanisms and immune microenvironment of NASH.

Characteristics of long-term survivors with malignant pleural mesothelioma

Pleural mesothelioma (PM) is a cancer with usually a dismal prognosis. However, long-term survivors do exist. Herein, we analyzed long-term survivors (>5 years after surgery) from high-volume centres around the world. This is a multicenter retrospective descriptive analysis of long-term survivors (overall survival ≥ 5 years from surgery) treated within a multimodality therapy approach including macroscopic complete resection. Overall survival was calculated with Kaplan Meier analysis and cases were matched by center and surgery year and compared with a control group of short-term survivors (<2 years) in a conditional logistic regression analysis. There were 276 long-term survivors, most were male (n=166, 63%) with a median age of 59 (range 21-83) years at time of diagnosis. The histology for 246 was epithelioid and non-epithelioid for 30 patients. The disease was on the right side in 58% of the patients. As of this analysis, 148 patients were dead, 104 were alive and 10 were lost to follow-up. Pathological tumor stages were: pT1 (n=50), pT2 (n=63), pT3 (n=90) or pT4 (n=16), pN0 (n=150), pN1 (n=20) and pN2 (n=39). The matched control dataset included 333 patients, 95 cases and 238 controls. Comparing short- with long-term survivors, there was moderate evidence that a low white blood cell (WBC) count before surgery was more often observed in long-term survivors. The data show that long-term survival in PM is possible in a subgroup of surgically treated patients; histological subtype and WBC count seem to be prognosticators for longer survival.

Wind energy system fault classification using deep CNN and improved PSO‐tuned extreme gradient boosting

AbstractIntelligent fault diagnosis for wind energy systems requires identifying unique characteristics to differentiate various fault types effectively, even when data discrepancy occurs due to the unpredictable and dynamic nature of its environment. This article addresses some of the challenges of fault classification in wind energy systems by proposing an integrated approach that combines deep learning features with a resampled supervisory control and data acquisition (SCADA) dataset. The methodology involves resampling the imbalanced SCADA dataset using synthetic minority oversampling technique (SMOTE) and near‐miss undersampling techniques, extracting deep learning features using deep convolutional neural network, and feeding them into an XGBoost (extreme gradient boosting) classifier with tuned parameters using adaptive elite‐particle swarm optimization (AEPSO). The effectiveness of the proposed method is demonstrated through validation conducted on a different imbalanced dataset showing superior performance metrics in terms of accuracy. Additionally, the study contributes to methodological advancements in wind turbine fault diagnosis by providing a rigorous framework for fault classification. It is confirmed that utilizing the extracted deep learning features into the resampled data can significantly affect the classification performance metrics. Furthermore, the proposed integrated approach shows significance for fault diagnosis enhancement in wind energy systems and advancing the field towards more efficient and reliable operation.

Leveraging Supervised Machine Learning Algorithms for System Suitability Testing of Mass Spectrometry Imaging Platforms.

Quality control and system suitability testing are vital protocols implemented to ensure the repeatability and reproducibility of data in mass spectrometry investigations. However, mass spectrometry imaging (MSI) analyses present added complexity since both chemical and spatial information are measured. Herein, we employ various machine learning algorithms and a novel quality control mixture to classify the working conditions of an MSI platform. Each algorithm was evaluated in terms of its performance on unseen data, validated with negative control data sets to rule out confounding variables or chance agreement, and utilized to determine the necessary sample size to achieve a high level of accurate classifications. In this work, a robust machine learning workflow was established where models could accurately classify the instrument condition as clean or compromised based on data metrics extracted from the analyzed quality control sample. This work highlights the power of machine learning to recognize complex patterns in MSI data and use those relationships to perform a system suitability test for MSI platforms.

Optimizing machine learning algorithms for diabetes data: A metaheuristic approach to balancing and tuning classifiers parameters

Diabetes mellitus poses a global health concern, prompting the development of machine learning algorithms designed to construct a model for the accurate classification of patients, enabling precise diagnoses and early-stage treatment. However, the efficacy of classifying diabetes patients through machine learning relies on datasets, often plagued by imbalance, leading to biased classification and inaccurate diagnoses. Previous research attempts, employing techniques like random sampling (under-sampling and oversampling) and the Synthetic Minority Oversampling Technique (SMOTE), have struggled to achieve optimally balanced datasets. Additionally, setting the best parameters for machine learning classifiers remains a challenging task. To address these issues, this research focuses on devising a methodological metaheuristic optimization, a machine learning algorithm tailored for diabetes data balancing, and classifier hyperparameter tuning. Leveraging Particle Swarm Optimization (PSO) algorithm for diabetes data balancing and a genetic algorithm to select the optimal architecture for various machine learning classifiers. The study compares the performance of the proposed metaheuristic data balancer and classifier architecture parameter tuner using classification metrics (F1 score, Average Precision–Recall (APR), AUC, and accuracy). The PSO balanced dataset emerges as the most effective in classifying diabetes, with an Average Percentage Improvement (API) in classification performance metrics: 20.78% accuracy, 16.79% area under the curve for receiver operating characteristics, and a significant 32.78% enhancement in APR. Moreover, the XGBOOST classifier trained with a genetic algorithm demonstrates minimal computational training time for the Centre for Disease Control and Prevention (CDC) diabetes dataset compared to the artificial neural network and random forest classifier. Notably, the imbalanced CDC diabetes dataset exhibits the least APR compared to random under-sampling and the PSO data balancing technique.

Clonal hematopoiesis in LGI1-antibody encephalitis.

Leucine-rich glioma-inactivated 1 (LGI1)-antibody encephalitis (LGI1e), the major form of autoimmune encephalitis (AE) presented with memory loss and faciobrachial dystonic seizure, commonly develops in aged population. Hematologic aging is often accompanied by clonal hematopoiesis (CH), a phenomenon in which specific mutations accumulate, potentially leading to autoimmune disorders or malignancies. Our research aimed to investigate the connection between clonal hematopoiesis of indeterminate potential (CHIP) and LGI1e. Peripheral blood samples from consecutive LGI1e patients were collected and analyzed for 24 clonal CHIP using targeted gene sequencing. The results were compared to a control dataset from an ethnically matched health care cohort. Patient characteristics were analyzed based on their CHIP status. A total of 52 LGI1e patients were enrolled for this study. Among them, three patients (5.8%) exhibited functional mutations in the ASXL1 gene, one of the CHIP-associated genes analyzed by targeted sequencing. This frequency was significantly higher compared to that of the control cohort (1%, p = 0.015). Nevertheless, the patients showed no difference in the clinical characteristics, laboratory results, and immunotherapy outcomes. LGI1e showed high frequency of ASXL1 functional mutation in the CHIP analysis, which may contribute to the underlying pathogenesis. Further research is needed to determine its direct role in the development of autoimmunity and disease progression.

Dialysis More Available Than Patient Education in Counties With High Diabetes Prevalence.

Poorly controlled diabetes is a principal cause of end stage renal disease (ESRD), generating an estimated 44% of new cases. Diabetes self-management education and support (DSMES) has been documented to reduce adverse outcomes such as ESRD. Helping patients better manage their condition could ultimately reduce ESRD prevalence. We compared the county-level availability of DSMES and dialysis as of November 2022 sorted by the estimated prevalence of diabetes among residents aged 18 years or older. The locations of DSMES programs and ESRD dialysis facilities were obtained from 2 professional organizations and the Centers for Medicare & Medicade Services. Estimated diabetes prevalence was obtained from the Centers for Disease Control and Prevention's PLACES data set. Counties were considered to have high diabetes prevalence if they fell into the top quartile for diabetes prevalence in 2019 (≥14.4% of adults). Analyses were conducted in 2023. DSMES was available in 41.0% of counties but in only 20.7% of counties with high diabetes prevalence versus 47.9% of low prevalence counties. Dialysis facilities were present in 59.2% of all counties, in 52.8% of all high diabetes prevalence counties, and in 61.4% of other counties. DSMES availability was linked to the presence of a hospital in the county, with only 6.3% of counties without a hospital offering the service. DSMES could play a role in reducing the prevalence of ESRD. Public health professionals need to be aware of the differing levels of local availability of this service and work to develop partnerships to provide DSMES in high-prevalence areas not currently served.

NETWORK ANALYSIS OF SINGLE-NUCLEOTIDE POLYMORPHISMS ASSOCIATED WITH ABERRANT INFLAMMATION IN TRAUMA PATIENTS SUGGESTS A ROLE FOR VESICLE-ASSOCIATED INFLAMMATORY PROGRAMS INVOLVING CD55.

Background: Critical illness stemming from severe traumatic injury is a leading cause of morbidity and mortality worldwide and involves the dysfunction of multiple organ systems, driven, at least in part, by dysregulated inflammation. We and others have shown a key role for genetic predisposition to dysregulated inflammation and downstream adverse critical illness outcomes. Recently, we demonstrated an association among genotypes at the single-nucleotide polymorphism (SNP) rs10404939 in LYPD4 , dysregulated systemic inflammation, and adverse clinical outcomes in a broad sample of ~1,000 critically ill patients. Methods: We sought to gain mechanistic insights into the role of LYPD4 in critical illness by bioinformatically analyzing potential interactions among rs10404939 and other SNPs. We analyzed a dataset of common (i.e., not rare) SNPs previously defined to be associated with genotype-specific, significantly dysregulated systemic inflammation trajectories in trauma patients, in comparison to a control dataset of common SNPs determined to exhibit an absence of genotype-specific inflammatory responses. Results: In the control dataset, this analysis implicated SNPs associated with phosphatidylinositol and various membrane transport proteins, but not LYPD4. In the patient subset with genotypically dysregulated inflammation, our analysis suggested the co-localization to lipid rafts of LYPD4 and the complement receptor CD55, as well as the neurally related CNTNAP2 and RIMS4. Segregation of trauma patients based on genotype of the CD55 SNP rs11117564 showed distinct trajectories of organ dysfunction and systemic inflammation despite similar demographics and injury characteristics. Conclusion: These analyses define novel interactions among SNPs that could enhance our understanding of the response to traumatic injury and critical illness.

Single-cell transcriptomic analysis reveals a decrease in the frequency of macrophage-RGS1high subsets in patients with osteoarticular tuberculosis

BackgroundCell subsets differentially modulate host immune responses to Mycobacterium tuberculosis (MTB) infection. However, the nature and functions of these subsets against osteoarticular tuberculosis (OTB) are unclear. Here, we aimed to understand the phenotypes and functions of immune cell subsets in patients with OTB using single-cell RNA sequencing (scRNA-Seq).MethodsPathological and healthy adjacent tissues were isolated from patients with OTB and subjected to scRNA-Seq. Unsupervised clustering of cells was performed based on gene expression profiles, and uniform manifold approximation and projection was used for clustering visualization.ResultsThirteen cell subsets were identified in OTB tissues. scRNA-seq datasets of patients and healthy controls (HCs) showed that infection changed the frequency of immune cell subsets in OTB tissues. Myeloid cell examination revealed nine subsets. The frequency of macrophage-RGS1high subsets decreased in OTB tissues; this increased MTB susceptibility in an SLC7A11/ferroptosis-dependent manner. Immunohistochemistry assays and flow cytometry for patients with OTB and osteoarticular bacterial infection (OBI) and HCs verified that the frequency of macrophage-RGS1high subset decreased in OTB tissues and blood samples, thereby distinguishing patients with OTB from HCs and patients with OBI.ConclusionThe macrophage-RGS1high subset levels were decreased in patients with OTB, and would be up-regulated after effective treatment. Therefore, the clinical significance of this study is to discover that macrophage-RGS1high subset may serve as a potential biomarker for OTB diagnosis and treatment efficacy monitoring.

Abstract We053: Altered inflammatory state and mitochondrial function identified by transcriptomics in paediatric congenital heart patients prior to surgical repair

Background: Congenital heart disease (CHD) remains the most common birth defect, with surgical intervention required in complex cases. In contrast to adult heart disease which disproportionately affects the left ventricle, complex CHD can be characterized by increased right ventricular (RV) pressure, leading to RV hypertrophy and eventually failure, with RV function known to be a major predictor in sustained cardiac health in these patients. Methods: RV samples were collected from discarded myocardial tissue from paediatric patients (0-16 years) undergoing their first cardiac surgery at Bristol Royal Hospital for Children or obtained from autopsy subjects. RNA sequencing and subsequent cellular deconvolution was conducted in CHD samples and integrated with publicly available control dataset. Cellular deconvolution, pathway network analysis and gene ontology were conducted and changes in expression of markers was confirmed via histological analysis. Results: Presence of CHD drove divergence of transcriptome with 511 differentially expressed genes (log fold change≥±2, padj≤0.05), with 82% shared homology. Known CHD genes; TBX5 , JAG1, NOTCH2 and NOTCH1 were reduced in CHD compared to controls (padj=2.24 E-07 , 1.38 E-27 , 7.76 E-31 , 9.11 E-7 ) . Conversely, mitochondrial dysfunction genes RPPH1 and RMPR (padj=4.67 E-132 , 2.23 E-107 ) were increased. Gene set enrichment identified mitochondrial dysfunctional pathways.Negative regulation of mitochondrial functions and metabolism was identified in functional network analysis. T helper cell markers CD4 and SELL were significantly downregulated (p&lt;0.0001, p=0.003). Cytotoxic T cell markers CD8a, LAGE3 and CD49a were significantly increased in CHD compared to control (p=0.0006, p&lt;0.0001, p=0.0118) as was proinflammatory marker Caspase1 (p=0.0055). Histological analysis confirmed an increase in CD45 and CD8 positive cellular bodies with the CHD tissue, which was absent in control. Deconvolution of bulk RNAseq data suggests a reduction in CD4+ T cells (p=0.0067) and an increase in CD8+ T cells (p=0.0223). Network analysis identified positive regulation of the immune system and cytokine signalling clusters within the inflammation functional network as were lymphocyte activation and leukocyte differentiation. Conclusions: Utilizing RV tissue from paediatric patients undergoing CHD cardiac surgery this study identifies dysfunctional mitochondrial pathways and an increase in inflammatory T cell presence prior to reparative surgery.

Coupled Conditional Neural Movement Primitives

Learning sensorimotor trajectories through flexible neural representations is fundamental for robots as it facilitates the building of motor skills as well as equipping them with the ability to represent the world as predictable temporal events. Recent advances in deep learning led to the development of powerful learning from demonstration (LfD) systems such as Conditional Neural Movement Primitives (CNMPs). CNMPs can robustly represent skills as movement distributions and allow them to be ‘recalled’ by conditioning the movement on a few observation points. In this study, we focus on improving CNMPs to achieve a higher resource economy by adopting a divide-and-conquer approach. We propose a novel neural architecture called Coupled CNMP (C-CNMP), that couples the latent spaces of a pair of CNMPs that splits a given sensorimotor trajectory into segments whose learning is undertaken by smaller sub-networks. Therefore, each sub-network needs to deal with a less complex trajectory making the learning less resource-hungry. With systematic simulations on a controlled trajectory data set, we show that the overhead brought by the coupling introduced in our model is well offset by the resource and performance gain obtained. To be specific, with CNMP model as the baseline, it is shown that the proposed model is able to learn to generate trajectories in the data set with a lower trajectory error measured as the mean absolute difference between the generated trajectory and the ground truth. Importantly, our model can perform well with relatively limited resources, i.e., with less number of neural network parameters compared to the baseline. To show that the findings from the controlled data set well-transfer to robot data, we use robot joint data in an LfD setting and compare the learning performance of the proposed model with the baseline model at equal complexity levels. The simulation experiments show that with also the robot joint data, the proposed model, C-CNMP, learns to generate the joint trajectories with significantly less error than the baseline model. Overall, our study improves the state of the art in sensorimotor trajectory learning and exemplifies how divide-and-conquer approaches can benefit deep learning architectures for resource economy.