Manifold Approximation Research Articles

K-hyperparameter tuning in high-dimensional genomics using joint optimization of deep differential evolutionary algorithm and unsupervised transfer learning from intelligent GenoUMAP embeddings

AbstractK-hyperparameter optimization in high-dimensional genomics remains a critical challenge, impacting the quality of clustering. Improved quality of clustering can enhance models for predicting patient outcomes and identifying personalized treatment plans. Subsequently, these enhanced models can facilitate the discovery of biomarkers, which can be essential for early diagnosis, prognosis, and treatment response in cancer research. Our paper addresses this challenge through a four-fold approach. Firstly, we empirically evaluate the k-hyperparameter optimization algorithms in genomics analysis using a correlation based feature selection method and a stratified k-fold cross-validation strategy. Secondly, we evaluate the performance of the best optimization algorithm in the first step using a variety of the dimensionality reduction methods applied for reducing the hyperparameter search spaces in genomics. Building on the two, we propose a novel algorithm for this optimization problem in the third step, employing a joint optimization of Deep-Differential-Evolutionary Algorithm and Unsupervised Transfer Learning from Intelligent GenoUMAP (Uniform Manifold Approximation and Projection). Finally, we compare it with the existing algorithms and validate its effectiveness. Our approach leverages UMAP pre-trained special autoencoder and integrates a deep-differential-evolutionary algorithm in tuning k. These choices are based on empirical analysis results. The novel algorithm balances population size for exploration and exploitation, helping to find diverse solutions and the global optimum. The learning rate balances iterations and convergence speed, leading to stable convergence towards the global optimum. UMAP’s superior performance, demonstrated by short whiskers and higher median values in the comparative analysis, informs its choice for training the special autoencoder in the new algorithm. The algorithm enhances clustering by balancing reconstruction accuracy, local structure preservation, and cluster compactness. The comprehensive loss function optimizes clustering quality, promotes hyperparameter diversity, and facilitates effective knowledge transfer. This algorithm’s multi-objective joint optimization makes it effective in genomics data analysis. The validation on this algorithm on three genomic datasets demonstrates superior clustering scores. Additionally, the convergence plots indicate relatively smoother curves and an excellent fitness landscape. These findings hold significant promise for advancing cancer research and computational genomics at large.

CCL4L2 participates in tendinopathy progression by promoting macrophage inflammatory responses: a single-cell analysis

BackgroundTendinopathy is very common in clinical practice, which is highly prevalent in athletes, sports enthusiasts and other people involved in high-load weight-bearing activities. Common types of tendinopathy include rotator cuff injury, Achilles tendinitis, tennis elbow and so on. Macrophages (Macs) are key immune cells in the pathogenesis of tendinopathy. In this study, CCL4L2+ M1-related signaling pathways were screened by combining single-cell RNA sequencing (scRNA-seq) to explore their significance in tendinopathy treatment.MethodsImmune cell populations were screened by Uniform Manifold Approximation and Projection (UMAP) downscaling, and Mac cell subsets were annotated using cell marker genes. The cellular communication mechanism between different cellular subsets such as Macs and tendon stem/progenitor cells (TSPCs) was demonstrated by cellular communication analysis. Based on cell marker genes of CCL4L2 + M1, Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment were performed to compare the expression differences in M1 and M2 between the Disease and Healthy groups. Associations between CCL4L2+ M1 and TSPCs were inferred by cell-cell communication analysis. The effects of CCL4L2 on Mac polarization and TSPCs were verified by enzyme-linked immunosorbent assay (ELISA) and real-time fluorescence quantitative PCR (qPCR).ResultsThe proportions of TSPCs, endothelial cells (ECs), smooth muscle cells (SMCs), and immune cells were significantly elevated in the Disease group. The proportion of M1 cells in the Disease group was higher than that in the Healthy group, while the proportion of M2 cells was lower than that in the Healthy group. M1 differentially expressed genes (DEGs) were mainly enriched to disease-related and immunoinflammation-related signaling pathways. Signaling intensities between M1 and TSPCs in pathways related to immunoinflammation and ischemic injury were significantly increased in the Disease group. The proportion of CCL4L2 + M1 in the Disease group was significantly higher than in the Healthy group, and communications between CCL4L2 + M1 and TSPCs varied significantly. Compared with the Control group, the expression levels of inflammatory cytokines were higher in the CCL4L2 group, and the expression levels of tendon differentiation markers (Egr1, Mkx, Scx, Type 1 collagen, Tnmd) were significantly down-regulated.ConclusionThe present study analyzed the heterogeneous alterations in the Healthy and Disease groups by scRNA-seq data and found that there was a significant inflammatory infiltrate in the Disease group with markedly increased Mac activity, which was associated with activation of the CCL4L2 + M1-associated signaling pathways. CCL4L2 promotes M1 polarization and inhibits TSPC differentiation through activating M1-related inflammatory signaling pathways. These findings contribute to a more comprehensive understanding of tendon injury progression and provide potential targets for tendinopathy treatment.

Variability in hot sub-luminous stars and binaries: Machine-learning analysis of Gaia DR3 multi-epoch photometry

Hot sub-luminous stars represent a population of stripped and evolved red giants that is located on the extreme horizontal branch. Since they exhibit a wide range of variability due to pulsations or binary interactions, it is crucial to unveil their intrinsic and extrinsic variability to understand the physical processes of their formation. In the Hertzsprung-Russell diagram, they overlap with interacting binaries such as cataclysmic variables (CVs). By leveraging the most recent clustering algorithm tools, we investigate the variability of 1,576 candidate hot subdwarf variables using comprehensive data from Gaia DR3 multi-epoch photometry and Transiting Exoplanet Survey Satellite (TESS) observations. We present a novel approach that uses the t-distributed stochastic neighbour embedding and the uniform manifold approximation and projection dimensionality reduction algorithms to facilitate the identification and classification of different populations of variable hot subdwarfs and CVs in a large dataset. In addition to the publicly available Gaia time-series statistics table, we adopted additional statistical features that enhanced the performance of the algorithms. The clustering results led to the identification of 85 new hot subdwarf variables based on Gaia and TESS light curves and of 108 new variables based on Gaia light curves alone, including reflection-effect systems, HW Vir, ellipsoidal variables, and high-amplitude pulsating variables. A significant number of known CVs (140) distinctively cluster in the 2D feature space among an additional 152 objects that we consider candidates for new CVs. This study paves the way for more efficient and comprehensive analyses of stellar variability from ground- and space-based observations, and for the application of machine-learning classifications of candidate variable stars in large surveys.

Quantitative evaluation of electrographic response to electroconvulsive therapy in super-refractory status epilepticus

ObjectiveElectroconvulsive therapy (ECT) has been occasionally applied as a treatment for super-refractory status epilepticus (SRSE). However, the effects of ECT on electrographic activity and related clinical outcomes are largely unknown. Here, we use quantitative approaches on electroencephalography (EEG) data to evaluate the neurophysiological influences of ECT and how they may relate to patient survival.MethodsThis was a single center study of adult patients who underwent bi-frontal ECT for treatment of SRSE between 2007 and 2021. Continuous scalp EEG data obtained before and after each ECT session was converted using a linelength transform and projected into low-dimensional space using complementary linear and non-linear dimensionality reduction techniques (principal component analysis and separately uniform manifold approximation). Differences between before versus after ECT were quantified using silhouette scores. Mixed effects models evaluated whether changes in mean scores were related to time (across sessions, and separately within sessions up to 1 h after treatment) and patient outcomes (survival).ResultsEight patients underwent ECT for SRSE, ranging from 3 to 12 sessions each. Four patients survived with chronic epilepsy and varying cognitive sequelae, and four died while hospitalized. Projecting EEG data into low-dimensional space revealed several sessions with visualizable differences in electrographic activity before versus after ECT treatment. Silhouette scores significantly increased as time elapsed up to 60 min after ECT and higher scores were related to survival, though there was no significant change in scores across successive ECT sessions.DiscussionECT is associated with changes in electrographic activity in certain patients, and such changes may be associated with survival, although our study was underpowered to detect more definitive treatment-related effects. Further quantitative neurophysiology studies, and potentially clinical trials, in larger groups of patients are warranted to study direct influences of ECT treatment given the devastating and often deadly outcomes of SRSE.

Explainable machine learning models for predicting the acute toxicity of pesticides to sheepshead minnow (Cyprinodon variegatus)

A quantitative structure–activity relationship (QSAR) study was conducted on 313 pesticides to predict their acute toxicity to Sheepshead minnow (Cyprinodon variegatus) by using DRAGON descriptors. Essentials accounting for a reliable model were all considered carefully, giving full consideration to the OECD (Organization for Economic Co-operation and Development) principles for QSAR acceptability in regulation during the model construction and assessment process. Nine variables were selected through the forward stepwise regression method and used as inputs to construct both linear and nonlinear models. The obtained models were validated internally and externally. Generally, machine learning-based methods, namely support vector machine (SVM), random forest (RF), and projection pursuit regression (PPR), perform better than the multiple linear regression (MLR) model. The statistical results (R2 = 0.682–0.933, Q2LOO = 0.604–0.659, Q2F1 = 0.740–0.796, CCC = 0.861–0.882) of the developed models show that they are robust, reliable, reproducible, accurate and predictive. Comparatively, the RF model performs best, giving predictive correlation coefficient Q2 of 0.814, root mean squared error (RMSE) of 0.658 and mean absolute error (MAE) of 0.534 for the test set, respectively. The RF model (as well as SVM and PPR models) was visualized and explained by using the SHapley Additive explanation (SHAP) analysis to enhance its transparency and credibility. In addition, the applicability domain (AD) range of the RF model was characterized by the Williams plot and the tree manifold approximation and projection (TMAP) technology was utilized to illustrate similarity and diversity of the entire data space, to assist in the analysis of the outliers. Activity cliff detection was investigated by using Arithmetic Residuals in K-groups Analysis (ARKA) descriptors. It was found that none of the pesticides was identified as an activity cliff in the training set or a potential prediction cliff in the test set. Therefore, the RF model fulfills each OECD principle in regulation for QSAR models. The research in this work will aid in the in silico QSAR prediction of the acute toxicity to Sheepshead minnow (Cyprinodon variegatus) for untested and new toxic pesticides and can also be extended to other studies.

Elevated frequencies of activated memory B cells in multiple sclerosis are reset to healthy control levels after B cell depletion with Ocrelizumab

In multiple sclerosis (MS) the B cell depleting drug ocrelizumab has shown high efficacy in reducing inflammatory activity. Its mechanism of action is unclear due to B cell subset complexity and unknown roles in pathogenesis. Here, we comprehensively phenotyped and quantitated peripheral blood B cell subsets before and after ocrelizumab infusion to gain insight into the fate of B cell subsets with pathogenic potential. Peripheral blood B cells were collected from treatment naïve patients at baseline and months one, three, and six following the first course of ocrelizumab treatment; at 6 months following the second treatment cycle; ~14 months following their last infusion; and from healthy controls. Flow cytometry combined with cluster analysis was used to track depletion and repletion of naïve, memory, and antibody secreting cells. By month one, naïve B cells were depleted, but a small subset of memory B cells were retained with no depletion of antibody secreting cells. Uniform manifold approximation and projection for dimension reduction analysis of flow cytometry data revealed two non-class switched naïve clusters and two class switched memory clusters. One class switched cluster was activated in MS patients but largely absent in healthy controls. Both memory B cell subsets underwent depletion after a single six-month course of ocrelizumab treatment after which their proportions were reset to heathy control levels. These observations suggest that activated class-switched memory B cells could serve as a biomarker of recent or ongoing MS disease activity to guide redosing.

Identification of the CD8+ T-cell exhaustion signature of hepatocellular carcinoma for the prediction of prognosis and immune microenvironment by integrated analysis of bulk- and single-cell RNA sequencing data.

Hepatocellular carcinoma (HCC) is a prevalent type of cancer with high incidence and mortality rates. It is the third most common cause of cancer-related deaths. CD8+ T cell exhaustion (TEX) is a progressive decline in T cell function due to sustained T cell receptor stimulation from continuous antigen exposure. Studies have shown that CD8+ TEX plays an important role in the anti-tumor immune process and is significantly correlated with patient prognosis. The aim of the research is to establish a reliable CD8+ TEX-based signature using single-cell RNA sequencing (scRNA-seq) and high-throughput RNA sequencing (RNA-seq), providing a new approach to evaluate HCC patient prognosis and immune microenvironment. The RNA-seq data of HCC patients were download from three different databases: The Cancer Genome Atlas (TCGA), the Gene Expression Omnibus (GEO), and the International Cancer Genome Consortium (ICGC). HCC's 10× scRNA data were acquired from GSE149614. Based on single-cell sequencing data, CD8+ TEX-related genes were identified using uniform manifold approximation and projection (UMAP) algorithm, singleR, and marker gene methods. Afterwards, we proceeded to construct CD8+ TEX signature using differential gene analysis, univariate Cox regression analysis, least absolute shrinkage and selection operator (LASSO) regression, and multivariate Cox regression analysis. We also validated the CD8+ TEX signature in GEO and ICGC external cohorts and investigated clinical characteristics, chemotherapy sensitivity, mutation landscape, functional analysis, and immune cell infiltration in different risk groups. The CD8+ TEX signature, consisting of 13 genes (HSPD1, UBB, DNAJB4, CALM1, LGALS3, BATF, COMMD3, IL7R, FDPS, DRAP1, RPS27L, PAPOLA, GPR171), was found to have a strong predictive effect on the prognosis of HCC. The Kaplan-Meier (KM) analysis showed that the overall survival (OS) rate of patients in the low-risk group was higher than that of patients in the high-risk group across different datasets and specific populations. The research findings suggested that the risk score was an independent predictor of HCC prognosis. The model based on clinical features and risk score has a strong predictive effect. We observed significant differences among various risk groups in terms of clinical characteristics, functional analysis, mutation landscape, chemotherapy sensitivity, and immune cell infiltration. We constructed a CD8+ TEX signature to predict the survival probability of patients with HCC. We also found that the model could predict the sensitivity of targeted drugs and immune cell infiltration, and the risk score was negatively correlated with CD8+ T cell infiltration. In summary, the CD8+ TEX signature of HCC was constructed for the prediction of prognosis and immune microenvironment by integrated analysis of bulk and scRNA-seq data.

Sparse attention regression network-based soil fertility prediction with UMMASO

The challenge of imbalanced soil nutrient datasets significantly hampers accurate predictions of soil fertility. To tackle this, a new method is suggested in this research, combining Uniform Manifold Approximation and Projection (UMAP) with Least Absolute Shrinkage and Selection Operator (LASSO). The main aim is to counter the impact of uneven data distribution and improve soil fertility models' predictive precision. The model introduced uses Sparse Attention Regression, effectively incorporating pertinent features from the imbalanced dataset. UMAP is utilised initially to reduce data complexity, unveiling hidden structures and essential patterns. Following this, LASSO is applied to refine features and enhance the model's interpretability. The experimental outcomes highlight the effectiveness of the UMAP and LASSO hybrid approach. The proposed model achieves outstanding performance metrics, reaching a predictive accuracy of 98 %, demonstrating its capability in accurate soil fertility predictions. It also showcases a Precision of 91.25 %, indicating its adeptness in accurately identifying fertile soil instances. The Recall metric stands at 90.90 %, emphasizing the model's ability to capture true positive cases effectively.

Compact DINO-ViT: Feature Reduction for Visual Transformer

Research has been ongoing for years to discover image features that enable their best classification. One of the latest developments in this area is the Self-Distillation with No Labels Vision Transformer—DINO-ViT features. However, even for a single image, their volume is significant. Therefore, for this article we proposed to substantially reduce their size, using two methods: Principal Component Analysis and Neighborhood Component Analysis. Our developed methods, PCA-DINO and NCA-DINO, showed a significant reduction in the volume of the features, often exceeding an order of magnitude while maintaining or slightly reducing the classification accuracy, which was confirmed by numerous experiments. Additionally, we evaluated the Uniform Manifold Approximation and Projection (UMAP) method, showing the superiority of the PCA and NCA approaches. Our experiments involving modifications to patch size, attention heads, and noise insertion in DINO-ViT demonstrated that both PCA-DINO and NCA-DINO exhibited reliable accuracy. While NCA-DINO is optimal for high-performance applications despite its higher computational cost, PCA-DINO offers a faster, more resource-efficient solution, depending on the application-specific requirements. The code for our method is available on GitHub.

Single‐cell analyses reveal impaired type B spermatogonia differentiation and meiotic entry in C‐Nap1‐null testes

AbstractSperm development is critical for male reproductive capability; any disruption during the process of spermatogenesis will result in male infertility. In this research, we used the C‐Nap1 encoded by the gene of Cep250 knockout mouse line as the model to evaluate the impact of absent C‐Nap1 on spermatogenesis. To investigate the interaction between C‐Nap1 and spermatogenesis, we utilized single‐cell RNA sequencing to analyze 10,332 C‐Nap1+/+ and 13,308 C‐Nap1−/− testicular cells. We identified five main cell types within seminiferous tubules, including spermatogonia, Sertoli cells, spermatogonia stem cells, Leydig cells, and spermatocytes. We found a critical reduction in testicular spermatogonia and spermatocytes in C‐Nap1‐null testes, compared to its C‐Nap1+/+ controls. By combining uniform manifold approximation and projection clustering and psedotime ordering, we distinguished five spermatogonial stages/subtypes, demonstrating that type B spermatogonia differentiation and meiotic initiation are impaired during C‐Nap1‐null spermatogenesis. Following gene ontology enrichment analysis, meiosis‐specific genes downregulated in the C‐Nap1−/− testicular cells were further verified by reverse transcription polymerase chain reaction (RT‐PCR). Based on the differential gene expression, certain downregulated genes such as Ctnnb1 and Aurka encoding C‐Nap1‐binding potential β‐Catenin and Aurka are encountered, which may account for defective type B spermatogonia differentiation and meiotic entry in C‐Nap1‐null testes.

Uniform manifold approximation and projection

Uniform manifold approximation and projection

Uniform manifold approximation and projection

Uniform manifold approximation and projection

Dimensionality Reduction for Data Analysis With Quantum Feature Learning

ABSTRACTTo improve data analysis and feature learning, this study compares the effectiveness of quantum dimensionality reduction (qDR) techniques to classical ones. In this study, we investigate several qDR techniques on a variety of datasets such as quantum Gaussian distribution adaptation (qGDA), quantum principal component analysis (qPCA), quantum linear discriminant analysis (qLDA), and quantum t‐SNE (qt‐SNE). The Olivetti Faces, Wine, Breast Cancer, Digits, and Iris are among the datasets used in this investigation. Through comparison evaluations against well‐established classical approaches, such as classical PCA (cPCA), classical LDA (cLDA), and classical GDA (cGDA), and using well‐established metrics like loss, fidelity, and processing time, the effectiveness of these techniques is assessed. The findings show that cPCA produced positive results with the lowest loss and highest fidelity when used on the Iris dataset. On the other hand, quantum uniform manifold approximation and projection (qUMAP) performs well and shows strong fidelity when tested against the Wine dataset, but ct‐SNE shows mediocre performance against the Digits dataset. Isomap and locally linear embedding (LLE) function differently depending on the dataset. Notably, LLE showed the largest loss and lowest fidelity on the Olivetti Faces dataset. The hypothesis testing findings showed that the qDR strategies did not significantly outperform the classical techniques in terms of maintaining pertinent information from quantum datasets. More specifically, the outcomes of paired t‐tests show that when it comes to the ability to capture complex patterns, there are no statistically significant differences between the cPCA and qPCA, the cLDA and qLDA, and the cGDA and qGDA. According to the findings of the assessments of mutual information (MI) and clustering accuracy, qPCA may be able to recognize patterns more clearly than standardized cPCA. Nevertheless, there is no discernible improvement between the qLDA and qGDA approaches and their classical counterparts.

Harmonizome 3.0: integrated knowledge about genes and proteins from diverse multi-omics resources.

By processing and abstracting diverse omics datasets into associations between genes and their attributes, the Harmonizome databaseenables researchers to explore and integrate knowledge about human genes from many central omics resources. Here, we introduce Harmonizome 3.0, a significant upgrade to the original Harmonizome database. The upgrade adds 26 datasets that contribute nearly 12 million associations between genes and various attribute types such as cells and tissues, diseases, and pathways.The upgrade has adatasetcrossing feature to identify gene modules that are shared across datasets. To further explain significantly high gene set overlap between dataset pairs, a large language model (LLM) composes a paragraph that speculates about the reasons behind the high overlap. The upgradealso adds more data formats and visualization options. Datasets aredownloadable as knowledge graph (KG) assertions and visualized withUniform Manifold Approximation and Projection (UMAP) plots. The KG assertions can be explored via a user interface that visualizes gene-attribute associations as ball-and-stick diagrams. Overall, Harmonizome 3.0 is a rich resource of processed omics datasets that are provided in several AI-ready formats.Harmonizome 3.0is available at https://maayanlab.cloud/Harmonizome/.

Integrative analysis of single-nucleus RNA sequencing and Mendelian randomization to explore novel risk genes for Alzheimer's disease.

In this study, we aimed to delineate cellular heterogeneity in Alzheimer's disease (AD) and identify genetic markers contributing to its pathogenesis using integrative analysis of single-nucleus RNA sequencing (sn-RNA-Seq) and Mendelian randomization (MR). The dorsolateral prefrontal cortex sn-RNA-Seq dataset (GSE243292) was sourced from the Gene Expression Omnibus (GEO) database. Data preprocessing was conducted using the Seurat R software package, employing principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) for cell clustering and annotation. MR analysis was used to identify instrumental variables from expression quantitative trait loci (eQTL) and GWAS data by applying inverse variance weighting (IVW), weighted median (WM) and MR-Egger methods. This was complemented by leave-one-out sensitivity analysis to validate the causal relationship on AD risk genes. We identified 23 distinct cell clusters, which were annotated into eight subgroups, including oligodendrocytes, oligodendrocyte precursors, astrocytes, macrophage cells, endothelial cells, glutamatergic neurons, neural stem cells, and neurons. Notably, the number of macrophages significantly increased in the AD group. Using genome-wide association study (GWAS) summaries and eQTL data, MR analysis identified causal relationships for 7 genes with significant impacts on AD risk. Among these genes, CACNA2D3, INPP5D, RBM47, and TBXAS1 were associated with a decreased risk of AD, whereas EPB41L2, MYO1F, and SSH2 were associated with an increased risk. A leave-one-out sensitivity analysis confirmed the robustness of these findings. Expression analysis revealed that these genes were variably expressed across different cell subgroups. Except for the CACNA2D3 gene, the other 6 genes showed increased expression levels in the macrophages, particularly EPB41L2 and SSH2. Our findings highlight the potential of specific genetic markers identified through integrative analysis of sn-RNA-Seq and MR in guiding the diagnosis and therapeutic strategies for Alzheimer's disease.

A comprehensive human embryo reference tool using single-cell RNA-sequencing data.

Stem cell-based embryo models offer unprecedented experimental tools for studying early human development. The usefulness of embryo models hinges on their molecular, cellular and structural fidelities to their in vivo counterparts. To authenticate human embryo models, single-cell RNA sequencing has been utilized for unbiased transcriptional profiling. However, an organized and integrated human single-cell RNA-sequencing dataset, serving as a universal reference for benchmarking human embryo models, remains unavailable. Here we developed such a reference through the integration of six published human datasets covering development from the zygote to the gastrula. Lineage annotations are contrasted and validated with available human and nonhuman primate datasets. Using stabilized Uniform Manifold Approximation and Projection, we constructed an early embryogenesis prediction tool, where query datasets can be projected on the reference and annotated with predicted cell identities. Using this reference tool, we examined published human embryo models, highlighting the risk of misannotation when relevant references are not utilized for benchmarking and authentication.

A machine learning approach to stratify patients with hypermobile Ehlers-Danlos syndrome/hypermobility spectrum disorders according to disorders of gut brain interaction, comorbidities and quality of life.

A high prevalence of disorders of gut-brain interaction (DGBI) exist in patients with hypermobile Ehlers-Danlos Syndrome (hEDS) and hypermobility spectrum disorders (HSD). However, it is unknown if clusters of hEDS/HSD patients exist which overlap with different DGBIs and whether this overlap influences presence of comorbidities and quality of life. We aimed to study these knowledge gaps. A prospectively collected hEDS/HSD cohort of 1044 individuals were studied. We undertook Uniform Manifold Approximation and Projection-enabled (UMAP) dimension reduction to create a representation of nonlinear interactions between hEDS/HSD and DGBIs, from which individuals were stratified into clusters. Somatization, Postural Tachycardia Syndrome (PoTS), autonomic symptoms, psychological factors and quality of life were statistically compared between clusters. The mean age of patients was 40 ± 13.2 years; 87.8% were female. Patients segregated into three clusters: Cluster 0 (n = 466): hEDS/HSD+ functional foregut disorders (FFD) + irritable bowel syndrome (IBS); Cluster 1 (n = 180): hEDS/HSD+ IBS and Cluster 2 (n = 337): hEDS/HSD alone. In cluster 0, we demonstrated increased somatization (p <0.0001), anxiety (p <0.0001), depression (p <0.0001), PoTS prevalence (p = 0.003), autonomic symptoms (p <0.0001) and reduced quality of life (p <0.0001) compared to cluster 2. Cluster 0 had greater comorbidity burden than cluster 1. Within hEDS/HSD, subgroups exist with a high prevalence of FFD and IBS. These subgroups have a higher prevalence of psychological disorders, dysautonomia and poorer quality of life compared with hEDS/HSD alone. Further research should focus on healthcare utilization, management and prognosis in hEDS/HSD and DGBI overlap.

Dimensionality reduction distills complex evolutionary relationships in seasonal influenza and SARS-CoV-2.

Public health researchers and practitioners commonly infer phylogenies from viral genome sequences to understand transmission dynamics and identify clusters of genetically-related samples. However, viruses that reassort or recombine violate phylogenetic assumptions and require more sophisticated methods. Even when phylogenies are appropriate, they can be unnecessary or difficult to interpret without specialty knowledge. For example, pairwise distances between sequences can be enough to identify clusters of related samples or assign new samples to existing phylogenetic clusters. In this work, we tested whether dimensionality reduction methods could capture known genetic groups within two human pathogenic viruses that cause substantial human morbidity and mortality and frequently reassort or recombine, respectively: seasonal influenza A/H3N2 and SARS-CoV-2. We applied principal component analysis, multidimensional scaling (MDS), t-distributed stochastic neighbor embedding (t-SNE), and uniform manifold approximation and projection to sequences with well-defined phylogenetic clades and either reassortment (H3N2) or recombination (SARS-CoV-2). For each low-dimensional embedding of sequences, we calculated the correlation between pairwise genetic and Euclidean distances in the embedding and applied a hierarchical clustering method to identify clusters in the embedding. We measured the accuracy of clusters compared to previously defined phylogenetic clades, reassortment clusters, or recombinant lineages. We found that MDS embeddings accurately represented pairwise genetic distances including the intermediate placement of recombinant SARS-CoV-2 lineages between parental lineages. Clusters from t-SNE embeddings accurately recapitulated known phylogenetic clades, H3N2 reassortment groups, and SARS-CoV-2 recombinant lineages. We show that simple statistical methods without a biological model can accurately represent known genetic relationships for relevant human pathogenic viruses. Our open source implementation of these methods for analysis of viral genome sequences can be easily applied when phylogenetic methods are either unnecessary or inappropriate.

Abstract 4135951: Single Nuclei RNA Sequencing Reveals Distinct Cellular Composition of Pericoronary Adipose Tissue

Introduction: Adipose tissue is a metabolically active organ with local paracrine interactions on adjacent tissues. Pericoronary adipose tissue (PcAT) is an emerging cardiovascular disease (CVD) risk factor and likely has paracrine effects on coronary vasculature that may promote coronary artery disease (CAD) development. PcAT may be functionally and clinically distinct from epicardial adipose tissue (EAT) given its proximity to the coronary artery adventitia. The cellular composition of PcAT and EAT and whether they have unique inflammatory cell signatures and functional expression at a single cell level has not been explored. Hypothesis: PcAT will have diverse cellular composition with heightened inflammatory cell content compared to EAT. Methods: Visual dissection of paired EAT not adjacent to an epicardial coronary artery and PcAT in direct apposition to an epicardial artery (left anterior descending artery) was performed from a single human heart immediately after explant (prior to heart transplantation). Nuclei were isolated using the Chromium Nuclei Isolation Kit (10x Genomics) with RNAse inhibitor. Cells were counted with a Cellometer K2 and Chromium Next GEM Single Cell libraries were prepared, followed by reverse transcription, PCR and NovaSeq 6000 platform sequencing. Raw sequence reads from single nuclei RNA-seq (snRNA-seq) were processed and after QC filtering, cells were clustered using Seurat. Clusters were annotated using singleR. Results: In PcAT and EAT, &gt;7000 cells were successfully extracted with 73755 and 48750 mean snRNA-seq reads per cell, respectively. snRNA-seq cellular composition analysis delineated a heightened inflammatory milieu and distinct cellular composition of human PcAT vs. EAT. A UMAP (Uniform Manifold Approximation and Projection) plot displays higher inflammatory cell infiltrate with a neutrophil predominance and lesser adipocyte composition in PcAT compared with EAT ( Figure 1) . Conclusions: We have generated snRNA-seq data from human PcAT and shown distinct cellular composition compared to EAT. Given the proximity of PcAT to coronary vasculature, these results are supportive of paracrine inflammatory effects and highlight adipose tissue biology in cardiovascular disease. Understanding this biology may identify novel therapeutic targets for CVD.

Abstract 4114276: Identification of the heterogeneity in the risk of heart failure in those with left bundle branch block pattern using artificial intelligence

Introduction: The association between left bundle branch block (LBBB) and heart failure is well-established. However, the lack of a practical risk assessment tool for patients with LBBB poses a clinical challenge to the management despite the heterogeneity of the LBBB population. Hypothesis: Artificial intelligence (AI) can extract features to stratify the risk of heart failure (HF) within the heterogeneity of LBBB. Methods: All 12-lead ECGs diagnosed with LBBB recorded from 1/1/2015 to 12/31/2019 were identified in our institution. ECG data were analyzed as 2D matrices with a shape of 12×2500. The Uniform Manifold Approximation and Projection (UMAP) was used to visualize the heterogeneity of LBBB ECG. Additionally, a 2-dimensional convolutional neural network model was trained to detect LBBB ECGs associated with past HF diagnosis. The model was then applied to ECGs from an external cohort of patients with LBBB but without a history of HF. Cumulative incidences of HF admission were compared by stratifying according to tertiles of the model output (low, intermediate and high AI score groups) Results: We identified 15,124 LBBB ECGs in our institution (training dataset) and 2,463 individuals with LBBB ECGs in the external cohort (external test dataset). The UMAP projection of the ECGs revealed distinct heterogeneity in the data ( Fig A ). This heterogeneity was not fully explained by patient demographics, ECG features, or institutions. When the external cohort was divided into 3 groups according to the tertile of AI prediction, patients with high AI scores had a higher risk of HF admission (high vs low AI score group: hazard ratio (HR), 2.10; 95% confidence interval (CI), 1.66-2.65; intermediate vs low AI score group: HR, 1.39; 95%CI, 1.08-1.78: log-rank p &lt;0.0001; Fig 1B ). This finding was unchanged after adjusting for multiple clinical characteristics and ECG parameters, including QRS duration. Conclusion: We observed significant heterogeneity in ECGs presenting LBBB, and the AI model excellently stratified the risk of heart failure admission in patients with LBBB from a single recording of 12-lead ECG, suggesting great potential for clinical utility in managing patients with LBBB.