ABSTRACT
The growing complexity of biological data has spurred the development of innovative computational techniques to extract meaningful information and uncover hidden patterns within vast datasets. Biological networks, such as gene regulatory networks and protein-protein interaction networks, hold critical insights into biological features' connections and functions. Integrating and analyzing high-dimensional data, particularly in gene expression studies, stands prominent among the challenges in deciphering these networks. Clustering methods play a crucial role in addressing these challenges, with spectral clustering emerging as a potent unsupervised technique considering intrinsic geometric structures. However, spectral clustering's user-defined cluster number can lead to inconsistent and sometimes orthogonal clustering regimes. We propose the Multi-layer Bundling (MLB) method to address this limitation, combining multiple prominent clustering regimes to offer a comprehensive data view. We call the outcome clusters "bundles". This approach refines clustering outcomes, unravels hierarchical organization, and identifies bridge elements mediating communication between network components. By layering clustering results, MLB provides a global-to-local view of biological feature clusters enabling insights into intricate biological systems. Furthermore, the method enhances bundle network predictions by integrating the bundle co-cluster matrix with the affinity matrix. The versatility of MLB extends beyond biological networks, making it applicable to various domains where understanding complex relationships and patterns is needed.
Subject(s)
Algorithms , Computational Biology , Gene Regulatory Networks , Mathematical Concepts , Protein Interaction Maps , Cluster Analysis , Humans , Models, Biological , Gene Expression Profiling/statistics & numerical data , Gene Expression Profiling/methodsABSTRACT
BACKGROUND: Sepsis from infection is a global health priority and clinical trials have failed to deliver effective therapeutic interventions. To address complicating heterogeneity in sepsis pathobiology, and improve outcomes, promising precision medicine approaches are helping identify disease endotypes, however, they require a more complete definition of sepsis subgroups. METHODS: Here, we use RNA sequencing from peripheral blood to interrogate the host response to sepsis from participants in a global observational study carried out in West Africa, Southeast Asia, and North America (N = 494). RESULTS: We identify four sepsis subtypes differentiated by 28-day mortality. A low mortality immunocompetent group is specified by features that describe the adaptive immune system. In contrast, the three high mortality groups show elevated clinical severity consistent with multiple organ dysfunction. The immunosuppressed group members show signs of a dysfunctional immune response, the acute-inflammation group is set apart by molecular features of the innate immune response, while the immunometabolic group is characterized by metabolic pathways such as heme biosynthesis. CONCLUSIONS: Our analysis reveals details of molecular endotypes in sepsis that support immunotherapeutic interventions and identifies biomarkers that predict outcomes in these groups.
Sepsis is a life-threatening multi-organ failure caused by the body's immune response to infection. Clinical symptoms of sepsis vary from one person to another likely due to differences in host factors, infecting pathogen, and comorbidities. This difference in clinical symptoms may contribute to the lack of effective interventions for sepsis. Therefore, approaches tailored to targeting groups of patients who present similarly are of great interest. This study analysed a large group of sepsis patients with diverse symptoms using laboratory markers and mathematical analysis. We report four patient groups that differ by risk of death and immune response profile. Targeting these defined groups with tailored interventions presents an exciting opportunity to improve the health outcomes of patients with sepsis.
ABSTRACT
Our limited understanding of the pathophysiological mechanisms that operate during sepsis is an obstacle to rational treatment and clinical trial design. There is a critical lack of data from low- and middle-income countries where the sepsis burden is increased which inhibits generalized strategies for therapeutic intervention. Here we perform RNA sequencing of whole blood to investigate longitudinal host response to sepsis in a Ghanaian cohort. Data dimensional reduction reveals dynamic gene expression patterns that describe cell type-specific molecular phenotypes including a dysregulated myeloid compartment shared between sepsis and COVID-19. The gene expression signatures reported here define a landscape of host response to sepsis that supports interventions via targeting immunophenotypes to improve outcomes.
Subject(s)
COVID-19 , Phenotype , Sepsis , Transcriptome , Humans , Sepsis/genetics , Sepsis/blood , Sepsis/immunology , COVID-19/immunology , COVID-19/genetics , COVID-19/blood , COVID-19/virology , Ghana/epidemiology , Male , Cohort Studies , SARS-CoV-2/immunology , SARS-CoV-2/genetics , Female , Adult , Middle Aged , Gene Expression Profiling , Sequence Analysis, RNAABSTRACT
The islets of Langerhans are critical endocrine micro-organs that secrete hormones regulating energy metabolism in animals. Insulin and glucagon, secreted by beta and alpha cells, respectively, are responsible for metabolic switching between fat and glucose utilization. Dysfunction in their secretion and/or counter-regulatory influence leads to diabetes. Debate in the field centers on the cytoarchitecture of islets, as the signaling that governs hormonal secretion depends on structural and functional factors, including electrical connectivity, innervation, vascularization, and physical proximity. Much effort has therefore been devoted to elucidating which architectural features are significant for function and how derangements in these features are correlated or causative for dysfunction, especially using quantitative network science or graph theory characterizations. Here, we ask if there are non-local features in islet cytoarchitecture, going beyond standard network statistics, that are relevant to islet function. An example is ring structures, or cycles, of α and δ cells surrounding ß cell clusters or the opposite, ß cells surrounding α and δ cells. These could appear in two-dimensional islet section images if a sphere consisting of one cell type surrounds a cluster of another cell type. To address these issues, we developed two independent computational approaches, geometric and topological, for such characterizations. For the latter, we introduce an application of topological data analysis to determine locations of topological features that are biologically significant. We show that both approaches, applied to a large collection of islet sections, are in complete agreement in the context both of developmental and diabetes-related changes in islet characteristics. The topological approach can be applied to three-dimensional imaging data for islets as well.
Subject(s)
Diabetes Mellitus , Insulin-Secreting Cells , Islets of Langerhans , Animals , Insulin/metabolism , Glucagon , Insulin-Secreting Cells/metabolism , Diabetes Mellitus/metabolismABSTRACT
OBJECTIVES: We evaluated the performance of commonly used sepsis screening tools across prospective sepsis cohorts in the USA, Cambodia and Ghana. DESIGN: Prospective cohort studies. SETTING AND PARTICIPANTS: From 2014 to 2021, participants with two or more SIRS (Systemic Inflammatory Response Syndrome) criteria and suspected infection were enrolled in emergency departments and medical wards at hospitals in Cambodia and Ghana and hospitalised participants with suspected infection were enrolled in the USA. Cox proportional hazards regression was performed, and Harrell's C-statistic calculated to determine 28-day mortality prediction performance of the quick Sequential Organ Failure Assessment (qSOFA) score ≥2, SIRS score ≥3, National Early Warning Score (NEWS) ≥5, Modified Early Warning Score (MEWS) ≥5 or Universal Vital Assessment (UVA) score ≥2. Screening tools were compared with baseline risk (age and sex) with the Wald test. RESULTS: The cohorts included 567 participants (42.9% women) including 187 participants from Kumasi, Ghana, 200 participants from Takeo, Cambodia and 180 participants from Durham, North Carolina in the USA. The pooled mortality was 16.4% at 28 days. The mortality prediction accuracy increased from baseline risk with the MEWS (C-statistic: 0.63, 95% CI 0.58 to 0.68; p=0.002), NEWS (C-statistic: 0.68; 95% CI 0.64 to 0.73; p<0.001), qSOFA (C-statistic: 0.70, 95% CI 0.64 to 0.75; p<0.001), UVA score (C-statistic: 0.73, 95% CI 0.69 to 0.78; p<0.001), but not with SIRS (0.60; 95% CI 0.54 to 0.65; p=0.13). Within individual cohorts, only the UVA score in Ghana performed better than baseline risk (C-statistic: 0.77; 95% CI 0.71 to 0.83; p<0.001). CONCLUSIONS: Among the cohorts, MEWS, NEWS, qSOFA and UVA scores performed better than baseline risk, largely driven by accuracy improvements in Ghana, while SIRS scores did not improve prognostication accuracy. Prognostication scores should be validated within the target population prior to clinical use.
Subject(s)
Sepsis , Adult , Female , Humans , Male , Prospective Studies , Sepsis/diagnosis , Systemic Inflammatory Response Syndrome/diagnosis , Cambodia , Cohort StudiesABSTRACT
The associations between clinical phenotypes of coronavirus disease 2019 (COVID-19) and the host inflammatory response during the transition from peak illness to convalescence are not yet well understood. Blood plasma samples were collected from 129 adult SARS-CoV-2 positive inpatient and outpatient participants between April 2020 and January 2021, in a multi-center prospective cohort study at 8 military hospitals across the United States. Plasma inflammatory protein biomarkers were measured in samples from 15 to 28 days post symptom onset. Topological Data Analysis (TDA) was used to identify patterns of inflammation, and associations with peak severity (outpatient, hospitalized, ICU admission or death), Charlson Comorbidity Index (CCI), and body mass index (BMI) were evaluated using logistic regression. The study population (n = 129, 33.3% female, median 41.3 years of age) included 77 outpatient, 31 inpatient, 16 ICU-level, and 5 fatal cases. Three distinct inflammatory biomarker clusters were identified and were associated with significant differences in peak disease severity (p < 0.001), age (p < 0.001), BMI (p < 0.001), and CCI (p = 0.001). Host-biomarker profiles stratified a heterogeneous population of COVID-19 patients during the transition from peak illness to convalescence, and these distinct inflammatory patterns were associated with comorbid disease and severe illness due to COVID-19.
Subject(s)
COVID-19 , Humans , Female , United States/epidemiology , Male , SARS-CoV-2 , Prospective Studies , Convalescence , Biomarkers , Phenotype , Severity of Illness Index , HospitalizationABSTRACT
BACKGROUND: Venous phlebotomy performed by trained personnel is critical for patient diagnosis and monitoring of chronic disease, but has limitations in resource-constrained settings, and represents an infection control challenge during outbreaks. Self-collection devices have the potential to shift phlebotomy closer to the point of care, supporting telemedicine strategies and virtual clinical trials. Here we assess a capillary blood micro-sampling device, the Tasso Serum Separator Tube (SST), for measuring blood protein levels in healthy subjects and non-hospitalized COVID-19 patients. METHODS: 57 healthy controls and 56 participants with mild/moderate COVID-19 were recruited at two U.S. military healthcare facilities. Healthy controls donated Tasso SST capillary serum, venous plasma and venous serum samples at multiple time points, while COVID-19 patients donated a single Tasso SST serum sample at enrolment. Concentrations of 17 protein inflammatory biomarkers were measured in all biospecimens by Ella multi-analyte immune-assay. RESULTS: Tasso SST serum protein measurements in healthy control subjects were highly reproducible, but their agreements with matched venous samples varied. Most of the selected proteins, including CRP, Ferritin, IL-6 and PCT, were well-correlated between Tasso SST and venous serum with little sample type bias, but concentrations of D-dimer, IL-1B and IL-1Ra were not. Self-collection at home with delayed sample processing was associated with significant concentrations differences for several analytes compared to supervised, in-clinic collection with rapid processing. Finally, Tasso SST serum protein concentrations were significantly elevated in in non-hospitalized COVID-19 patients compared with healthy controls. CONCLUSIONS: Self-collection of capillary blood with micro-sampling devices provides an attractive alternative to routine phlebotomy. However, concentrations of certain analytes may differ significantly from those in venous samples, and factors including user proficiency, temperature control and time lags between specimen collection and processing need to be considered for their effect on sample quality and reproducibility.
Subject(s)
COVID-19 , Blood Proteins , Blood Specimen Collection , COVID-19/diagnosis , Healthy Volunteers , Humans , Reproducibility of Results , Specimen HandlingABSTRACT
Sepsis is a life-threatening condition and understanding the disease pathophysiology through the use of host immune response biomarkers is critical for patient stratification. Lack of accurate sepsis endotyping impedes clinicians from making timely decisions alongside insufficiencies in appropriate sepsis management. This work aims to demonstrate the potential feasibility of a data-driven validation model for supporting clinical decisions to predict sepsis host-immune response. Herein, we used a machine learning approach to determine the predictive potential of identifying sepsis host immune response for patient stratification by combining multiple biomarker measurements from a single plasma sample. Results were obtained using the following cytokines and chemokines IL-6, IL-8, IL-10, IP-10 and TRAIL where the test dataset was 70%. Supervised machine learning algorithm naïve Bayes and decision tree algorithm showed good accuracy of 96.64% and 94.64%. These promising findings indicate the proposed AI approach could be a valuable testing resource for promoting clinical decision making.
Subject(s)
Algorithms , Biomarkers/analysis , Machine Learning , Sepsis/diagnosis , Bayes Theorem , Case-Control Studies , Clinical Decision-Making , Humans , Reproducibility of ResultsABSTRACT
Plasma glucose in mammals is regulated by hormones secreted by the islets of Langerhans embedded in the exocrine pancreas. Islets consist of endocrine cells, primarily α, ß, and δ cells, which secrete glucagon, insulin, and somatostatin, respectively. ß cells form irregular locally connected clusters within islets that act in concert to secrete insulin upon glucose stimulation. Varying demands and available nutrients during development produce changes in the local connectivity of ß cells in an islet. We showed in earlier work that graph theory provides a framework for the quantification of the seemingly stochastic cyto-architecture of ß cells in an islet. To quantify the dynamics of endocrine connectivity during development requires a framework for characterizing changes in the probability distribution on the space of possible graphs, essentially a Fokker-Planck formalism on graphs. With large-scale imaging data for hundreds of thousands of islets containing millions of cells from human specimens, we show that this dynamics can be determined quantitatively. Requiring that rearrangement and cell addition processes match the observed dynamic developmental changes in quantitative topological graph characteristics strongly constrained possible processes. Our results suggest that there is a transient shift in preferred connectivity for ß cells between 1-35 weeks and 12-24 months.
Subject(s)
Islets of Langerhans/cytology , Islets of Langerhans/growth & development , Cell Count , Child, Preschool , Computer Graphics , Computer Simulation , Glucagon/analysis , Glucagon/metabolism , Humans , Infant , Infant, Newborn , Insulin/analysis , Insulin/metabolism , Insulin-Secreting Cells/cytology , Insulin-Secreting Cells/metabolism , Islets of Langerhans/metabolism , Models, Biological , Stochastic ProcessesABSTRACT
Pancreatic islets of Langerhans consist of endocrine cells, primarily α, ß and δ cells, which secrete glucagon, insulin, and somatostatin, respectively, to regulate plasma glucose. ß cells form irregular locally connected clusters within islets that act in concert to secrete insulin upon glucose stimulation. Due to the central functional significance of this local connectivity in the placement of ß cells in an islet, it is important to characterize it quantitatively. However, quantification of the seemingly stochastic cytoarchitecture of ß cells in an islet requires mathematical methods that can capture topological connectivity in the entire ß-cell population in an islet. Graph theory provides such a framework. Using large-scale imaging data for thousands of islets containing hundreds of thousands of cells in human organ donor pancreata, we show that quantitative graph characteristics differ between control and type 2 diabetic islets. Further insight into the processes that shape and maintain this architecture is obtained by formulating a stochastic theory of ß-cell rearrangement in whole islets, just as the normal equilibrium distribution of the Ornstein-Uhlenbeck process can be viewed as the result of the interplay between a random walk and a linear restoring force. Requiring that rearrangements maintain the observed quantitative topological graph characteristics strongly constrained possible processes. Our results suggest that ß-cell rearrangement is dependent on its connectivity in order to maintain an optimal cluster size in both normal and T2D islets.
Subject(s)
Computational Biology/methods , Insulin-Secreting Cells/cytology , Islets of Langerhans/cytology , Adult , Aged , Aged, 80 and over , Algorithms , Diabetes Mellitus, Type 2/physiopathology , Female , Humans , Male , Middle Aged , Stochastic ProcessesABSTRACT
A nucleotide sequence 35 base pairs long can take 1,180,591,620,717,411,303,424 possible values. An example of systems biology datasets, protein binding microarrays, contain activity data from about 40,000 such sequences. The discrepancy between the number of possible configurations and the available activities is enormous. Thus, albeit that systems biology datasets are large in absolute terms, they oftentimes require methods developed for rare events due to the combinatorial increase in the number of possible configurations of biological systems. A plethora of techniques for handling large datasets, such as Empirical Bayes, or rare events, such as importance sampling, have been developed in the literature, but these cannot always be simultaneously utilized. Here we introduce a principled approach to Empirical Bayes based on importance sampling, information theory, and theoretical physics in the general context of sequence phenotype model induction. We present the analytical calculations that underlie our approach. We demonstrate the computational efficiency of the approach on concrete examples, and demonstrate its efficacy by applying the theory to publicly available protein binding microarray transcription factor datasets and to data on synthetic cAMP-regulated enhancer sequences. As further demonstrations, we find transcription factor binding motifs, predict the activity of new sequences and extract the locations of transcription factor binding sites. In summary, we present a novel method that is efficient (requiring minimal computational time and reasonable amounts of memory), has high predictive power that is comparable with that of models with hundreds of parameters, and has a limited number of optimized parameters, proportional to the sequence length.
Subject(s)
Base Sequence , Bayes Theorem , Entropy , Algorithms , Binding Sites , Empirical Research , Systems BiologyABSTRACT
Previous studies describing the symptomatic onset of type 1 diabetes (T1D) and rate of beta-cell loss (C-peptide) support the notion that childhood onset T1D exhibits more severe beta-cell depletion compared to adult onset T1D. To test this notion, we performed whole pancreas analyses in two T1D cases, one of childhood onset (7-year old, onset at 1.5-year) along with an adult onset case (43-year old with onset at 27-year). Both cases were matched for age and gender with control subjects. Striking regional differences in beta-cell loss were observed in both T1D cases, with severity of loss in the order of tail > body > head regions. In contrast, pancreatic alpha- and delta-cell mass was similar in controls and T1D patients. In the childhood onset T1D case, no intra-islet beta-cells were detected while in the adult onset case, beta-cell containing islets were found, exclusively in the head region. In the latter case, considerable numbers of small cellular clusters negative for three major endocrine hormones were observed, in islets with or without beta-cells. Ultrastructural analysis suggests these cells correspond to degenerating beta-cells, with empty granular membranes and abnormal morphology of nuclei with intranuclear pseudo-inclusions, adjacent to healthy alpha- and delta-cells. These results support a hypothesis that during T1D development in childhood, beta-cells are more susceptible to autoimmune destruction or immune attack is more severe, while beta-cell death in the adult onset T1D may be more protracted and incomplete. In addition, T1D may be associated with the formation of "empty" beta-cells, an interesting population of cells that may represent a key facet to the disorder's pathogenesis.
Subject(s)
Diabetes Mellitus, Type 1/pathology , Insulin-Secreting Cells/pathology , Adult , Age of Onset , Child , Female , Glucagon-Secreting Cells/pathology , Glucagon-Secreting Cells/ultrastructure , Humans , Immunohistochemistry , Infant , Insulin-Secreting Cells/ultrastructure , Male , Pancreas/pathology , Pancreatic Function Tests , Somatostatin-Secreting Cells/pathology , Somatostatin-Secreting Cells/ultrastructureABSTRACT
The mechanism for cortical folding pattern formation is not fully understood. Current models represent scenarios that describe pattern formation through local interactions, and one recent model is the intermediate progenitor model. The intermediate progenitor (IP) model describes a local chemically driven scenario, where an increase in intermediate progenitor cells in the subventricular zone correlates to gyral formation. Here we present a mathematical model that uses features of the IP model and further captures global characteristics of cortical pattern formation. A prolate spheroidal surface is used to approximate the ventricular zone. Prolate spheroidal harmonics are applied to a Turing reaction-diffusion system, providing a chemically based framework for cortical folding. Our model reveals a direct correlation between pattern formation and the size and shape of the lateral ventricle. Additionally, placement and directionality of sulci and the relationship between domain scaling and cortical pattern elaboration are explained. The significance of this model is that it elucidates the consistency of cortical patterns among individuals within a species and addresses inter-species variability based on global characteristics and provides a critical piece to the puzzle of cortical pattern formation.