Causal Probabilistic Networks Research Articles

Probabilistic Causal Network Modeling of Southern Hemisphere Jet Subseasonal to Seasonal Predictability

Abstract Skillful prediction of the Southern Hemisphere (SH) eddy-driven jet is crucial for representation of mid-to-high-latitude SH climate variability. In the austral spring-to-summer months, the jet and the stratospheric polar vortex variabilities are strongly coupled. Since the vortex is more predictable and influenced by long-lead drivers 1 month or more ahead, the stratosphere is considered a promising pathway for improving forecasts in the region on subseasonal to seasonal (S2S) time scales. However, a quantification of this predictability has been lacking, as most modeling studies address only one of the several interacting drivers at a time, while statistical analyses quantify association but not skill. This methodological gap is addressed through a knowledge-driven probabilistic causal network approach, quantified with seasonal ensemble hindcast data. The approach enables to quantify the jet’s long-range predictability arising from known late-winter drivers, namely, El Niño–Southern Oscillation (ENSO), Indian Ocean dipole (IOD), upward wave activity flux, and polar night jet oscillation, mediated by the vortex variability in spring. Network-based predictions confirm the vortex as determinant for skillful jet predictions, both for the jet’s poleward shift in late spring and its equatorward shift in early summer. ENSO, IOD, late-winter wave activity flux, and polar night jet oscillation only provide moderate prediction skill to the vortex. This points to early spring submonthly variability as important for determining the vortex state leading up to its breakdown, creating a predictability bottleneck for the jet. The method developed here offers a new avenue to quantify the predictability provided by multiple, interacting drivers on S2S time scales. Significance Statement Predictions of the Southern Hemisphere midlatitude jet stream are crucial for skillful forecasts of the austral mid-to-high latitudes. Several oceanic and atmospheric phenomena could, if better represented in models, improve spring-to-summer jet predictions on subseasonal to seasonal time scales. However, the combined potential skill arising from the inclusion of such phenomena has not been quantified. This study does so by using a probabilistic causal network model, representing the connections between those drivers and the jet with conditional probabilities, trained on large sets of model data. The stratospheric polar vortex is confirmed as crucial predictor of jet variability but is itself hard to predict a month in advance due to submonthly variability, creating a predictability bottleneck for the jet.

Abstract 6544: Targeting resistance mechanisms to AR-targeted therapy in prostate cancer through inhibition of CREBBP/EP300

Abstract Standard initial treatment for men with metastatic prostate cancer involves targeting the androgen receptor (AR) with androgen deprivation therapy or androgen-receptor signaling inhibitors. Resistance eventually develops, such as through AR-activating alterations, leaving few targeted options for patients. Here, we utilize data from 4,509 prostate cancer patients with longitudinal clinical and DNA/RNA sequencing data to elucidate mechanisms of resistance and identify potential therapeutic vulnerabilities in resistant disease. By constructing comprehensive regulatory networks, termed hybrid networks, built across different disease stages, combining features from WGCNA, causal mediation analysis, probabilistic causal network inference, and public knowledge graphs, we uncover subnetworks strongly associated with disease progression and resistance. These subnetworks reflect biological processes for cell cycle, E2F, DNA repair, MYC, and AR, comprehensively identifying broad transcriptional changes that drive disease progression and resistance. Selected subnetwork scores, derived from RNA expression, were significantly associated with outcomes in both the hormone-sensitive and castration-resistant settings, demonstrating that the subnetworks capture early and late resistance. The hybrid networks identified the transcriptional co-activators CREBBP and EP300 as key modulators of these subnetworks, suggesting that targeting CREBBP/EP300 may be effective in treating both early and late-stage disease. Prostate cancer cell lines were treated with FT-6876, a small molecule inhibitor of CREBBP/EP300, and growth inhibition was measured in both enzalutamide sensitive and resistant cell lines, covering a wide range of resistance mechanisms. The selected subnetworks were enriched for genes down-regulated following treatment with FT-6876, suggesting that targeting CREBBP/P300 addresses multiple resistance-related pathways. Differential changes in protein abundance and protein acetylation between enzalutamide sensitive and resistant cell lines identified changes in fatty acid metabolism as a common, downstream event in response to treatment with FT-6876, suggesting an additional mechanism for response to FT-6876. Together, our integrated theoretical and experimental analyses support the hypothesis that targeting CREBBP/EP300 may provide therapeutic benefits to prostate cancer patients. Further, the subnetworks identified may serve as mechanistically-informed biomarkers for identifying patients who may benefit from inhibition of CREBBP/EP300. Citation Format: Bonnie V. Dougherty, Beatrice C. Thomas, Alice M. Walsh, Veronica Calvo, Kathryne E. Ware, Andrew J. Armstrong, Jens Renstrup, Jason A. Somarelli, Eric E. Schadt. Targeting resistance mechanisms to AR-targeted therapy in prostate cancer through inhibition of CREBBP/EP300 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6544.

Integrated study of systemic and local airway transcriptomes in asthma reveals causal mediation of systemic effects by airway key drivers

BackgroundSystemic and local profiles have each been associated with asthma, but parsing causal relationships between system-wide and airway-specific processes can be challenging. We sought to investigate systemic and airway processes in asthma and their causal relationships.MethodsThree hundred forty-one participants with persistent asthma and non-asthmatic controls were recruited and underwent peripheral blood mononuclear cell (PBMC) collection and nasal brushing. Transcriptome-wide RNA sequencing of the PBMC and nasal samples and a series of analyses were then performed using a discovery and independent test set approach at each step to ensure rigor. Analytic steps included differential expression analyses, coexpression and probabilistic causal (Bayesian) network constructions, key driver analyses, and causal mediation models.ResultsAmong the 341 participants, the median age was 13 years (IQR = 10–16), 164 (48%) were female, and 200 (58.7%) had persistent asthma with mean Asthma Control Test (ACT) score 16.6 (SD = 4.2). PBMC genes associated with asthma were enriched in co-expression modules for NK cell-mediated cytotoxicity (fold enrichment = 4.5, FDR = 6.47 × 10−32) and interleukin production (fold enrichment = 2.0, FDR = 1.01 × 10−15). Probabilistic causal network and key driver analyses identified NK cell granule protein (NKG7, fold change = 22.7, FDR = 1.02 × 10−31) and perforin (PRF1, fold change = 14.9, FDR = 1.31 × 10−22) as key drivers predicted to causally regulate PBMC asthma modules. Nasal genes associated with asthma were enriched in the tricarboxylic acid (TCA) cycle module (fold enrichment = 7.5 FDR = 5.09 × 10−107), with network analyses identifying G3BP stress granule assembly factor 1 (G3BP1, fold change = 9.1 FDR = 2.77 × 10−5) and InaD-like protein (INADL, fold change = 5.3 FDR = 2.98 × 10−9) as nasal key drivers. Causal mediation analyses revealed that associations between PBMC key drivers and asthma are causally mediated by nasal key drivers (FDR = 0.0076 to 0.015).ConclusionsIntegrated study of the systemic and airway transcriptomes in a well-phenotyped asthma cohort identified causal key drivers of asthma among PBMC and nasal transcripts. Associations between PBMC key drivers and asthma are causally mediated by nasal key drivers.

Predicting sepsis onset using a machine learned causal probabilistic network algorithm based on electronic health records data

Sepsis is a leading cause of mortality and early identification improves survival. With increasing digitalization of health care data automated sepsis prediction models hold promise to aid in prompt recognition. Most previous studies have focused on the intensive care unit (ICU) setting. Yet only a small proportion of sepsis develops in the ICU and there is an apparent clinical benefit to identify patients earlier in the disease trajectory. In this cohort of 82,852 hospital admissions and 8038 sepsis episodes classified according to the Sepsis-3 criteria, we demonstrate that a machine learned score can predict sepsis onset within 48 h using sparse routine electronic health record data outside the ICU. Our score was based on a causal probabilistic network model—SepsisFinder—which has similarities with clinical reasoning. A prediction was generated hourly on all admissions, providing a new variable was registered. Compared to the National Early Warning Score (NEWS2), which is an established method to identify sepsis, the SepsisFinder triggered earlier and had a higher area under receiver operating characteristic curve (AUROC) (0.950 vs. 0.872), as well as area under precision-recall curve (APR) (0.189 vs. 0.149). A machine learning comparator based on a gradient-boosting decision tree model had similar AUROC (0.949) and higher APR (0.239) than SepsisFinder but triggered later than both NEWS2 and SepsisFinder. The precision of SepsisFinder increased if screening was restricted to the earlier admission period and in episodes with bloodstream infection. Furthermore, the SepsisFinder signaled median 5.5 h prior to antibiotic administration. Identifying a high-risk population with this method could be used to tailor clinical interventions and improve patient care.

Bacteremia Prediction With Prognostic Scores and a Causal Probabilistic Network - A Cohort Study of Emergency Department Patients

BackgroundPhysicians tend to overestimate patients’ pretest probability of having bacteremia. The low yield of blood cultures and contaminants is associated with significant financial cost, as well as increased length of stay and unnecessary antibiotic treatment. ObjectiveThis study examined the abilities of the National Early Warning Score (NEWS), the Quick Sequential Organ Failure Assessment (qSOFA), the Modified Sequential Organ Failure Assessment (mSOFA), and two versions of the causal probabilistic network, SepsisFinder™ (SF) to predict bacteremia in adult emergency department (ED) patients. MethodsThis cohort study included adult ED patients from a large urban, academic tertiary hospital, with blood cultures obtained within 24 h of admission between 2016 and 2017. The outcome measure was true bacteremia. NEWS, qSOFA, mSOFA, and the two versions of SF score were calculated for all patients based on the first available full set of vital signs within 2 h and laboratory values within 6 h after drawing the blood cultures. Area under the receiver operating characteristic curve (AUROC) was calculated for each scoring system. ResultsThe study included 3106 ED patients, of which 199 (6.4%) patients had true bacteremia. The AUROCs for prediction of bacteremia were: NEWS = 0.65, qSOFA = 0.60, SF I = 0.65, mSOFA = 0.71, and SF II = 0.80. ConclusionsScoring systems using only vital signs, NEWS, and SF I showed moderate abilities in predicting bacteremia, whereas qSOFA performed poorly. Scoring systems using both vital signs and laboratory values, mSOFA and especially SF II, showed good abilities in predicting bacteremia.

Machine-Learning Model for Mortality Prediction in Patients With Community-Acquired Pneumonia: Development and Validation Study

Artificial intelligence tools and techniques such as machine learning (ML) are increasingly seen as a suitable manner in which to increase the prediction capacity of currently available clinical tools, including prognostic scores. However, studies evaluating the efficacy of ML methods in enhancing the predictive capacity of existing scores for community-acquired pneumonia (CAP) are limited. We aimed to apply and validate a causal probabilistic network (CPN) model to predict mortality in patients with CAP. Is a CPN model able to predict mortality in patients with CAP better than the commonly used severity scores? This was a derivation-validation retrospective study conducted in two Spanish university hospitals. The ability of a CPN designed to predict mortality in sepsis (SepsisFinder [SeF]), and adapted for CAP (SeF-ML), to predict 30-day mortality was assessed and compared with other scoring systems (Pneumonia Severity Index [PSI], Sequential Organ Failure Assessment [SOFA], quick Sequential Organ Failure Assessment [qSOFA], and CURB-65 criteria [confusion, urea, respiratory rate, BP, age≥ 65 years]). The SeF models are proprietary software. Differences between receiver operating characteristic curves were assessed by the DeLong method for correlated receiver operating characteristic curves. The derivation cohort comprised 4,531 patients, and the validation cohort consisted of 1,034 patients. In the derivation cohort, the areas under the curve (AUCs) of SeF-ML, CURB-65, SOFA, PSI, and qSOFA were 0.801, 0.759, 0.671, 0.799, and 0.642, respectively, for 30-day mortality prediction. In the validation study, the AUC of SeF-ML was 0.826, concordant with the AUC (0.801) in the derivation data (P= .51). The AUC of SeF-ML was significantly higher than those of CURB-65 (0.764; P= .03) and qSOFA (0.729, P= .005). However, it did not differ significantly from those of PSI (0.830; P= .92) and SOFA (0.771; P= .14). SeF-ML shows potential for improving mortality prediction among patients with CAP, using structured health data. Additional external validation studies should be conducted to support generalizability.

Downregulation of exhausted cytotoxic T cells in gene expression networks of multisystem inflammatory syndrome in children

Multisystem inflammatory syndrome in children (MIS-C) presents with fever, inflammation and pathology of multiple organs in individuals under 21 years of age in the weeks following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Although an autoimmune pathogenesis has been proposed, the genes, pathways and cell types causal to this new disease remain unknown. Here we perform RNA sequencing of blood from patients with MIS-C and controls to find disease-associated genes clustered in a co-expression module annotated to CD56dimCD57+ natural killer (NK) cells and exhausted CD8+ T cells. A similar transcriptome signature is replicated in an independent cohort of Kawasaki disease (KD), the related condition after which MIS-C was initially named. Probing a probabilistic causal network previously constructed from over 1,000 blood transcriptomes both validates the structure of this module and reveals nine key regulators, including TBX21, a central coordinator of exhausted CD8+ T cell differentiation. Together, this unbiased, transcriptome-wide survey implicates downregulation of NK cells and cytotoxic T cell exhaustion in the pathogenesis of MIS-C.

Machine-Learning Model for Mortality Prediction in Patients with Community-Acquired Pneumonia: Development and Validation Study

Background: Pneumonia is the leading infectious cause of mortality globally. Artificial intelligence tools and techniques such as machine learning (ML) are increasingly seen as a suitable manner to increase the prediction capacity of currently available tools in infectious diseases. However, studies evaluating the efficacy of ML methods in enhancing the predictive capacity of existing scores for community-acquired pneumonia (CAP) are limited. We aimed to apply a causal probabilistic network (CPN) model (SepsisFinder) to predict 30-day mortality in patients with CAP. As secondary objective, we pursued validating the ML model using a large cohort of CAP patients. Methods: This was a derivation-validation retrospective study to test the CPN model in its prediction of 30-day mortality in patients hospitalised with CAP. The derivation cohort comprised all consecutive adults with CAP admitted between 2003 and 2016 to a university hospital in Barcelona, Spain. For the validation cohort, we included all consecutive patients with CAP admitted to a university hospital in Valencia, Spain between 2012 and 2018. Findings: The derivation cohort comprised 4,531 patients whilst the validation cohort had 1,034 patients. In the derivation cohort, the AUC of SepsisFinder, CURB-65, SOFA, PSI and qSOFA were 0.801, 0.759, 0.671, 0.799 and 0.642, respectively, for 30-day mortality prediction. In the validation study, the AUC of SepsisFinder was 0.826, concordantly with the AUC (0.801) in the derivation data (p=0.51). The AUC of SepisFinder was significantly higher than those of CURB-65 (0.764, p=0.03) and qSOFA (0.729, p=0.005). However, it did not differ significantly from PSI (0.830, p=0.92) and SOFA (0.771, p=0.14). Interpretation: Sepsisfinder shows potential for improving mortality prediction amongst patients with CAP using structured health data. Additional external validation studies should be conducted to support generalisability. Funding: CIBER de Enfermedades Respiratorias (CIBERES CB06/06/0028), and by 2009 Support to Research Groups of Catalonia 911, IDIBAPS. Declaration of Interest: Dr Cilloniz is the recipient of the SEPAR fellowship 2018, and a grant from the Fondo de Investigacion Sanitaria (PI19/00207). LW is employed on a consulting basis and is a shareholder of Treat Systems ApS. MM is CEO and a shareholder of Treat Systems ApS. Treat Systems ApS is a manufacturer of medical device software for decision support in the field of infectious diseases and clinical microbiology, including TREAT-Lab. All other authors have nothing to declare. Ethical Approval: Institutional approval was provided by the Ethics Committee of the University Hospital La Fe of Valencia (XXXXXXX) and the Comite Etic d’Investigacio Clinica of the Hospital Clinic of Barcelona (2011/0219), which waived the need for informed consent. All data set were anonymously analyzed, and the study was performed following current recommendation of the Declaration of Helsinki.

1640-P: Common and Ethnic-Specific Molecular Networks and Drivers of Insulin Resistance in African and European Ancestry Individuals

Our understanding of how genetic variants influence molecular regulatory networks in glucose responsive tissues, and in turn affect cellular processes related to insulin resistance (IR) in different ethnic groups, remains incomplete. We employ a network modelling approach to integrate measures of insulin sensitivity, transcriptomic data in the adipose and muscle tissues, and genetic data from African Americans (AAs) in the AAGMEx cohort (N=256). Data from European ancestry (EA) individuals from AREA (N=99) and METSIM (N=770) cohorts are used as contrast. Multiscale Embedded GEne coexpression Network Analysis (MEGENA) in AAs identifies 167 and 63 transcript subnetworks (modules) enriched for Matsuda index and SI-correlated genes in adipose and muscle, respectively. Among these IR-associated modules, 11 in adipose and 6 in muscle, are enriched for the cis-eSNP genes. Bayesian probabilistic causal networks (BN) are constructed and subsequently used to identify key network drivers. Integration of MEGENA and BN in AAs identifies 99 and 39 key regulators for IR-associated adipose and muscle modules, respectively. Among the IR-associated modules in AAs, 23 in adipose and 5 in muscle are strongly preserved in the respective networks in the EA cohorts. Top 3 IR-associated preserved adipose modules are enriched for immune system, mitochondria and hormonal response and these modules are potentially regulated by KYNU, LDHD and ENO1, respectively. We identify 50 and 24 IR-associated but AA-specific modules (not preserved in EAs) in adipose and muscle, respectively. Conversely, we find 28 IR-associated but EA-specific modules in adipose. Ethnic group specific and shared modules reveal novel molecular mechanisms for IR. Enrichment of cis-eSNP genes in the network neighborhoods of key drivers suggest that regulatory SNPs may configure the networks by modulating transcript expression in adipose and muscle tissues, and are likely causal determinants for IR. Disclosure P. Xu: None. M. Wang: None. N.K. Sharma: None. M. Comeau: None. M. Civelek: None. C.D. Langefeld: None. B. Zhang: None. S.K. Das: None. Funding National Institutes of Health (R01DK118243, R01DK090111)

A Network Analysis of Multiple Myeloma Related Gene Signatures.

Multiple myeloma (MM) is the second most prevalent hematological cancer. MM is a complex and heterogeneous disease, and thus, it is essential to leverage omics data from large MM cohorts to understand the molecular mechanisms underlying MM tumorigenesis, progression, and drug responses, which may aid in the development of better treatments. In this study, we analyzed gene expression, copy number variation, and clinical data from the Multiple Myeloma Research Consortium (MMRC) dataset and constructed a multiple myeloma molecular causal network (M3CN). The M3CN was used to unify eight prognostic gene signatures in the literature that shared very few genes between them, resulting in a prognostic subnetwork of the M3CN, consisting of 178 genes that were enriched for genes involved in cell cycle (fold enrichment = 8.4, p value = 6.1 × 10−26). The M3CN was further used to characterize immunomodulators and proteasome inhibitors for MM, demonstrating the pleiotropic effects of these drugs, with drug-response signature genes enriched across multiple M3CN subnetworks. Network analyses indicated potential links between these drug-response subnetworks and the prognostic subnetwork. To elucidate the structure of these important MM subnetworks, we identified putative key regulators predicted to modulate the state of these subnetworks. Finally, to assess the predictive power of our network-based models, we stratified MM patients in an independent cohort, the MMRF-CoMMpass study, based on the prognostic subnetwork, and compared the performance of this subnetwork against other signatures in the literature. We show that the M3CN-derived prognostic subnetwork achieved the best separation between different risk groups in terms of log-rank test p-values and hazard ratios. In summary, this work demonstrates the power of a probabilistic causal network approach to understanding molecular mechanisms underlying the different MM signatures.

Integrative approach to sporadic Alzheimer\u2019s disease:\xa0deficiency of TYROBP\xa0in cerebral A\u03b2 amyloidosis mouse normalizes clinical phenotype and complement subnetwork molecular pathology without reducing A\u03b2 burden

Integrative gene network approaches enable new avenues of exploration that implicate causal genes in sporadic late-onset Alzheimer’s disease (LOAD) pathogenesis, thereby offering novel insights for drug-discovery programs. We previously constructed a probabilistic causal network model of sporadic LOAD and identified TYROBP/DAP12, encoding a microglial transmembrane signaling polypeptide and direct adapter of TREM2, as the most robust key driver gene in the network. Here, we show that absence of TYROBP/DAP12 in a mouse model of AD-type cerebral Aβ amyloidosis (APPKM670/671NL/PSEN1Δexon9) recapitulates the expected network characteristics by normalizing the transcriptome of APP/PSEN1 mice and repressing the induction of genes involved in the switch from homeostatic microglia to disease-associated microglia (DAM), including Trem2, complement (C1qa, C1qb, C1qc, and Itgax), Clec7a and Cst7. Importantly, we show that constitutive absence of TYROBP/DAP12 in the amyloidosis mouse model prevented appearance of the electrophysiological and learning behavior alterations associated with the phenotype of APPKM670/671NL/PSEN1Δexon9 mice. Our results suggest that TYROBP/DAP12 could represent a novel therapeutic target to slow, arrest, or prevent the development of sporadic LOAD. These data establish that the network pathology observed in postmortem human LOAD brain can be faithfully recapitulated in the brain of a genetically manipulated mouse. These data also validate our multiscale gene networks by demonstrating how the networks intersect with the standard neuropathological features of LOAD.

EXPLORING THE REPRODUCIBILITY OF PROBABILISTIC CAUSAL MOLECULAR NETWORK MODELS.

Network reconstruction algorithms are increasingly being employed in biomedical and life sciences research to integrate large-scale, high-dimensional data informing on living systems. One particular class of probabilistic causal networks being applied to model the complexity and causal structure of biological data is Bayesian networks (BNs). BNs provide an elegant mathematical framework for not only inferring causal relationships among many different molecular and higher order phenotypes, but also for incorporating highly diverse priors that provide an efficient path for incorporating existing knowledge. While significant methodological developments have broadly enabled the application of BNs to generate and validate meaningful biological hypotheses, the reproducibility of BNs in this context has not been systematically explored. In this study, we aim to determine the criteria for generating reproducible BNs in the context of transcription-based regulatory networks. We utilize two unique tissues from independent datasets, whole blood from the GTEx Consortium and liver from the Stockholm-Tartu Atherosclerosis Reverse Network Engineering Team (STARNET) study. We evaluated the reproducibility of the BNs by creating networks on data subsampled at different levels from each cohort and comparing these networks to the BNs constructed using the complete data. To help validate our results, we used simulated networks at varying sample sizes. Our study indicates that reproducibility of BNs in biological research is an issue worthy of further consideration, especially in light of the many publications that now employ findings from such constructs without appropriate attention paid to reproducibility. We find that while edge-to-edge reproducibility is strongly dependent on sample size, identification of more highly connected key driver nodes in BNs can be carried out with high confidence across a range of sample sizes.

Reasoning about causal relationships: Inferences on causal networks.

Over the last decade, a normative framework for making causal inferences, Bayesian Probabilistic Causal Networks, has come to dominate psychological studies of inference based on causal relationships. The following causal networks-[X→Y→Z, X←Y→Z, X→Y←Z]-supply answers for questions like, "Suppose both X and Y occur, what is the probability Z occurs?" or "Suppose you intervene and make Y occur, what is the probability Z occurs?" In this review, we provide a tutorial for how normatively to calculate these inferences. Then, we systematically detail the results of behavioral studies comparing human qualitative and quantitative judgments to the normative calculations for many network structures and for several types of inferences on those networks. Overall, when the normative calculations imply that an inference should increase, judgments usually go up; when calculations imply a decrease, judgments usually go down. However, 2 systematic deviations appear. First, people's inferences violate the Markov assumption. For example, when inferring Z from the structure X→Y→Z, people think that X is relevant even when Y completely mediates the relationship between X and Z. Second, even when people's inferences are directionally consistent with the normative calculations, they are often not as sensitive to the parameters and the structure of the network as they should be. We conclude with a discussion of productive directions for future research.

An integrated analysis of the marine social–ecological system of the Strait of Georgia, Canada, over the past four decades, and development of a regime shift index

A Driver–Pressure–State–Impact conceptual model of the Strait of Georgia, Canada, was developed as a framework to characterise the low frequency temporal (regime-like) patterns of variability in this system since 1970, and to identify suitable indicators of these changes. Thirty-seven time series, including both bio-physical (“natural”) and human variables, were assembled with sufficiently few missing years to permit multivariate statistical analyses. Principal component analyses identified regime-like shifts that correspond to those generally accepted for the NE Pacific (late 1970s; mid-1990s), plus an additional shift in the mid-1980s. Redundancy analyses identified significant relationships in which the natural and human explanatory driver/pressure variables accounted for 72% of the variability of the state/impact response variables. Partitioning this variance among the natural and human drivers/pressures attributed 4% to the natural drivers/pressures alone and 16% to the human drivers/pressures alone, but 53% to the interactions between natural and human drivers and pressures. Both sets of driver and pressure variables (natural and human) are therefore necessary to account for the patterns of variability of the state and impact variables in the Strait of Georgia. Redundancy analysis was also used to identify a reduced set of driver/pressure variables which accounted for a large proportion of the variability in the state and impact response variables: sea surface temperature, wind speed, the North Pacific Gyre Oscillation index, human population surrounding the Strait of Georgia, recreational fishing effort, and the number of hatchery releases of Chinook salmon (Oncorhynchus tshawytscha) into the Strait. A structural change analysis was used to identify significant changes in these six explanatory variables, and to identify regime shift years (including their 95% confidence intervals). A regime shift index for the Strait of Georgia was then constructed using a probabilistic causal network model. This model predicted the observed regime shift behaviour with a success rate of >70%. These six explanatory variables are proposed as leading indicators for low frequency changes (regime shifts) in the Strait of Georgia marine social–ecological system for use within a broader monitoring program.

Can computed tomography classifications of chronic obstructive pulmonary disease be identified using Bayesian networks and clinical data?

Diagnosis and classification of chronic obstructive pulmonary disease (COPD) may be seen as difficult. Causal reasoning can be used to relate clinical measurements with radiological representation of COPD phenotypes airways disease and emphysema. In this paper a causal probabilistic network was constructed that uses clinically available measurements to classify patients suffering from COPD into the main phenotypes airways disease and emphysema. The network grades the severity of disease and for emphysematous COPD, the type of bullae and its location central or peripheral. In four patient cases the network was shown to reach the same conclusion as was gained from the patients’ High Resolution Computed Tomography (HRCT) scans. These were: airways disease, emphysema with central small bullae, emphysema with central large bullae, and emphysema with peripheral bullae. The approach may be promising in targeting HRCT in COPD patients, assessing phenotypes of the disease and monitoring its progression using clinical data.

CR1 and the “Vanishing Amyloid” Hypothesis of Alzheimer’s Disease

CR1 and the “Vanishing Amyloid” Hypothesis of Alzheimer’s Disease

Stitching together Multiple Data Dimensions Reveals Interacting Metabolomic and Transcriptomic Networks That Modulate Cell Regulation

Author SummaryIt is now possible to score variations in DNA across whole genomes, RNA levels and alternative isoforms, metabolite levels, protein levels and protein state information, protein–protein interactions, and protein–DNA interactions, in a comprehensive fashion in populations of individuals. Interactions among these molecular entities define the complex web of biological processes that give rise to all higher order phenotypes, including disease. The development of analytical approaches that simultaneously integrate different dimensions of data is essential if we are to extract the meaning from large-scale data to elucidate the complexity of living systems. Here, we use a novel Bayesian network reconstruction algorithm that simultaneously integrates DNA variation, RNA levels, metabolite levels, protein–protein interaction data, protein–DNA binding data, and protein–small-molecule interaction data to construct molecular networks in yeast. We demonstrate that these networks can be used to infer causal relationships among genes, enabling the identification of novel genes that modulate cellular regulation. We show that our network predictions either recapitulate known biology or can be prospectively validated, demonstrating a high degree of accuracy in the predicted network.

Algorithm for Controlling Mechanical Properties of Hot Rolled Steels Using Bayesian Network Model

The work presents a control method for on-line adjustment of mechanical properties of as rolled steels produced at hot strip mills. The key idea of the method is a probabilistic causal (Bayesian) network which represents in a form of a directed acyclic graph the joint probability distribution of mechanical properties, chemical composition and temperature–strain parameters acting during hot rolling. As a slab moves along the mill the distribution is used for continuous recalculating the posterior probability of all mechanical properties conditioned by chemical composition and all other process parameters which become known to the moment of recalculating. Finally, when a strip is just before the finishing group we evaluate the probability distribution of finishing rolling temperature and coil temperature given the strip has the target mechanical properties. It generates new setups for these temperatures A pilot version of the method has been just implemented at CSP–line at Vyksa, Russia, United Metallurgical Company’s steel production site The adjustment is realized through appropriate correction of finish rolling temperature or/and coiling temperature setups of the mill automatic control system after the last chemical analysis of the current heat is made at the start of casting. Only “cautious” corrections of the temperatures are permissible so far (deviation from predefined level not more than ±30 degrees for each temperature) and the main aim of them is to set off the influence of chemistry variations on mechanical properties scatter of a given steel grade. The results of using the algorithm show that even these limited but interconnected actions reduce approximately twice the standard deviation of the mechanical properties inside a steel grade.

Cognition Evolutionary Computation for System-of-systems Architecture Development

The evolving nature of system-of-systems requirements and corresponding architecture, and the complex causal relationship between the architecture of system-of-systems and its emergent behavior make the problem of system-of-systems architecture development a great challenge. As a tentative effort in meeting this special challenge, a new evolutionary computation paradigm–named Cognition Evolutionary Computation(CEC) is proposed, which models the creative cognition process of divergent and convergent thinking, adaptation and innovation, that drives the co-evolution of problem space, knowledge space and solution space. The optimization algorithm for CEC uses causal probabilistic network as the knowledge representation mechanism. A theoretical framework for CEC based system-of-system architecture generation, evaluation and optimization is discussed.

A model for diagnosis of pulmonary infections in solid-organ transplant recipients

Background Opportunistic pulmonary infections are a major cause of morbidity and mortality among solid organ transplant recipients. The diagnosis of these infections is challenging because of the broad spectrum of bacteria, fungi and viruses affecting these patients and the lack of specific signs and symptoms. Treatment directed at the offending organism started as soon as possible improves survival. Objective To develop a decision support system for the diagnosis of pulmonary infections in solid-organ transplant recipients. The model's goal is to improve the accuracy of the diagnosis and thus the appropriateness of empirical treatment. Design The model is built using a Bayesian network (also known as causal probabilistic network). The network is based on pathogen segments which are the main building blocks of the model. Segments share common risk factors, such as time after transplantation, latent infections of donor/recipient and organ transplanted. The segments are linked at symptoms, signs and diagnostic tests common to all pathogens. The outputs of the model are predicted probabilities of infectious pathogens. To populate the model with data we have mainly abstracted data from the literature, using a systematic approach. The structure of the model and its adaptation for decision support will be presented. Evaluation The first evaluation phase assessed the model's diagnosis in a series of 20 representative cases of opportunistic infections. A match between the case's diagnosis and the model's prediction was achieved in 17/20 of cases. The next evaluation phase will consist of a prospective observational study comparing the accuracy of the model's diagnosis vs. that of the physician within 24 h of episode onset, as compared with a gold-standard diagnosis ascribed to the patients at the end of the infectious episode by two independent experts. Data for this phase are currently collected prospectively.