Evaluating robustness of a generalized linear model when applied to electronic health record data accessed using an Open API.

The Integrated Clinical and Environmental Exposures Service (ICEES) provides open regulatory-compliant access to clinical data, including electronic health record data, that have been integrated with environmental exposures data. While ICEES has been validated in the context of an asthma use case and several other use cases, the regulatory constraints on the ICEES open application programming interface (OpenAPI) result in data loss when using the service for multivariate analysis. In this study, we investigated the robustness of the ICEES OpenAPI through a comparative analysis, in which we applied a generalized linear model (GLM) to the OpenAPI data and the constraint-free source data to examine factors predictive of asthma exacerbations. Consistent with previous studies, we found that the main predictors identified by both analyses were sex, prednisone, race, obesity, and airborne particulate exposure. Comparison of GLM model fit revealed that data loss impacts model quality, but only with select interaction terms. We conclude that the ICEES OpenAPI supports multivariate analysis, albeit with potential data loss that users should be aware of.

Open Application of Statistical and Machine Learning Models to Explore the Impact of Environmental Exposures on Health and Disease: An Asthma Use Case.

ICEES (Integrated Clinical and Environmental Exposures Service) provides a disease-agnostic, regulatory-compliant approach for openly exposing and analyzing clinical data that have been integrated at the patient level with environmental exposures data. ICEES is equipped with basic features to support exploratory analysis using statistical approaches, such as bivariate chi-square tests. We recently developed a method for using ICEES to generate multivariate tables for subsequent application of machine learning and statistical models. The objective of the present study was to use this approach to identify predictors of asthma exacerbations through the application of three multivariate methods: conditional random forest, conditional tree, and generalized linear model. Among seven potential predictor variables, we found five to be of significant importance using both conditional random forest and conditional tree: prednisone, race, airborne particulate exposure, obesity, and sex. The conditional tree method additionally identified several significant two-way and three-way interactions among the same variables. When we applied a generalized linear model, we identified four significant predictor variables, namely prednisone, race, airborne particulate exposure, and obesity. When ranked in order by effect size, the results were in agreement with the results from the conditional random forest and conditional tree methods as well as the published literature. Our results suggest that the open multivariate analytic capabilities provided by ICEES are valid in the context of an asthma use case and likely will have broad value in advancing open research in environmental and public health.

An approach for open multivariate analysis of integrated clinical and environmental exposures data.

The Integrated Clinical and Environmental Exposures Service (ICEES) provides regulatory-compliant open access to sensitive patient data that have been integrated with public exposures data. ICEES was designed initially to support dynamic cohort creation and bivariate contingency tests. The objective of the present study was to develop an open approach to support multivariate analyses using existing ICEES functionalities and abiding by all regulatory constraints. We first developed an open approach for generating a multivariate table that maintains contingencies between clinical and environmental variables using programmatic calls to the open ICEES application programming interface. We then applied the approach to data on a large cohort (N = 22,365) of patients with asthma or related conditions and generated an eight-feature table. Due to regulatory constraints, data loss was incurred with the incorporation of each successive feature variable, from a starting sample size of N = 22,365 to a final sample size of N = 4,556 (20.4%), but data loss was < 10% until the addition of the final two feature variables. We then applied a generalized linear model to the subsequent dataset and focused on the impact of seven select feature variables on asthma exacerbations, defined as annual emergency department or inpatient visits for respiratory issues. We identified five feature variables-sex, race, obesity, prednisone, and airborne particulate exposure-as significant predictors of asthma exacerbations. We discuss the advantages and disadvantages of ICEES open multivariate analysis and conclude that, despite limitations, ICEES can provide a valuable resource for open multivariate analysis and can serve as an exemplar for regulatory-compliant informatic solutions to open patient data, with capabilities to explore the impact of environmental exposures on health outcomes.

R/DWD: distance-weighted discrimination for classification, visualization and batch adjustment.

UNLABELLED: R/DWD is an extensible package for classification. It is built based on a recently developed powerful classification method called distance weighted discrimination (DWD). DWD is related to, and has been shown to be superior to, the support vector machine in situations that are fundamental to bioinformatics, such as very high dimensional data. DWD has proven to be very useful for several fundamental bioinformatics tasks, including classification, data visualization and removal of biases, such as batch effects. Earlier DWD implementations, however, relied on Matlab, which is not free and requires a license. The major contribution of the R/DWD package is an implementation that is completely in R and thus can be used without any requirements for licensing or software purchase. In addition, R/DWD also provides efficient solvers for second-order-cone-programming and quadratic programming. AVAILABILITY AND IMPLEMENTATION: The package is freely available from cran.r-project.org.

Flow-based combinatorial antibody profiling: an integrated approach to cell characterization.

BD FACS™ CAP (CAP = combinatorial antibody profile) is a screening tool for rapid characterization of human cell surface protein expression profiles using semi-automated high-throughput flow cytometry. The current configuration consists of 229 directly conjugated antibodies arrayed in a 96-well plate as three-color cocktails, which enables the characterization of each of the 229 individual surface markers. Each individual cell type of interest is analyzed on the 96-well screening plates and the data are acquired on a flow cytometer equipped with a high-throughput sampler. The expression level of each marker for each cell type is then calculated using semiautomated custom flow cytometry software. The process of characterizing these surface markers in a highly efficient manner using BD FACS™ CAP is enabled by automated liquid handling for staining, automated flow cytometry for data acquisition, and standardized algorithms for automated data analysis.

Analysis of High-Throughput Flow Cytometry Data Using plateCore.

Flow cytometry (FCM) software packages from R/Bioconductor, such as flowCore and flowViz, serve as an open platform for development of new analysis tools and methods. We created plateCore, a new package that extends the functionality in these core packages to enable automated negative control-based gating and make the processing and analysis of plate-based data sets from high-throughput FCM screening experiments easier. plateCore was used to analyze data from a BD FACS CAP screening experiment where five Peripheral Blood Mononucleocyte Cell (PBMC) samples were assayed for 189 different human cell surface markers. This same data set was also manually analyzed by a cytometry expert using the FlowJo data analysis software package (TreeStar, USA). We show that the expression values for markers characterized using the automated approach in plateCore are in good agreement with those from FlowJo, and that using plateCore allows for more reproducible analyses of FCM screening data.

flowCore: a Bioconductor package for high throughput flow cytometry.

BACKGROUND: Recent advances in automation technologies have enabled the use of flow cytometry for high throughput screening, generating large complex data sets often in clinical trials or drug discovery settings. However, data management and data analysis methods have not advanced sufficiently far from the initial small-scale studies to support modeling in the presence of multiple covariates. RESULTS: We developed a set of flexible open source computational tools in the R package flowCore to facilitate the analysis of these complex data. A key component of which is having suitable data structures that support the application of similar operations to a collection of samples or a clinical cohort. In addition, our software constitutes a shared and extensible research platform that enables collaboration between bioinformaticians, computer scientists, statisticians, biologists and clinicians. This platform will foster the development of novel analytic methods for flow cytometry. CONCLUSION: The software has been applied in the analysis of various data sets and its data structures have proven to be highly efficient in capturing and organizing the analytic work flow. Finally, a number of additional Bioconductor packages successfully build on the infrastructure provided by flowCore, open new avenues for flow data analysis.

Functional T cell responses to tumor antigens in breast cancer patients have a distinct phenotype and cytokine signature.

The overall prevalence with which endogenous tumor Ags induce host T cell responses is unclear. Even when such responses are detected, they do not usually result in spontaneous remission of the cancer. We hypothesized that this might be associated with a predominant phenotype and/or cytokine profile of tumor-specific responses that is different from protective T cell responses to other chronic Ags, such as CMV. We detected significant T cell responses to CEA, HER-2/neu, and/or MAGE-A3 in 17 of 21 breast cancer patients naive to immunotherapy. The pattern of T cell cytokines produced in response to tumor-associated Ags (TAAs) in breast cancer patients was significantly different from that produced in response to CMV or influenza in the same patients. Specifically, there was a higher proportion of IL-2-producing CD8(+) T cells, and a lower proportion of IFN-gamma-producing CD4(+) and/or CD8(+) T cells responding to TAAs compared with CMV or influenza Ags. Finally, the phenotype of TAA-responsive CD8(+) T cells in breast cancer patients was almost completely CD28(+)CD45RA(-) (memory phenotype). CMV-responsive CD8(+) T cells in the same patients were broadly distributed among phenotypes, and contained a high proportion of terminal effector cells (CD27(-)CD28(-)CD45RA(+)) that were absent in the TAA responses. Taken together, these results suggest that TAA-responsive T cells are induced in breast cancer patients, but those T cells are phenotypically and functionally different from CMV- or influenza-responsive T cells. Immunotherapies directed against TAAs may need to alter these T cell signatures to be effective.

Data standards for flow cytometry.

Flow cytometry (FCM) is an analytical tool widely used for cancer and HIV/AIDS research, and treatment, stem cell manipulation and detecting microorganisms in environmental samples. Current data standards do not capture the full scope of FCM experiments and there is a demand for software tools that can assist in the exploration and analysis of large FCM datasets. We are implementing a standardized approach to capturing, analyzing, and disseminating FCM data that will facilitate both more complex analyses and analysis of datasets that could not previously be efficiently studied. Initial work has focused on developing a community-based guideline for recording and reporting the details of FCM experiments. Open source software tools that implement this standard are being created, with an emphasis on facilitating reproducible and extensible data analyses. As well, tools for electronic collaboration will assist the integrated access and comprehension of experiments to empower users to collaborate on FCM analyses. This coordinated, joint development of bioinformatics standards and software tools for FCM data analysis has the potential to greatly facilitate both basic and clinical research--impacting a notably diverse range of medical and environmental research areas.

IL-2 production correlates with effector cell differentiation in HIV-specific CD8+ T cells.

BACKGROUND: Diminished IL-2 production and lack of effector differentiation have been reported for HIV-specific T cells. In this study, we examined the prevalence of these phenomena using 8-color cytokine flow cytometry, and tested the hypothesis that these two findings were causally related. We analyzed cytokine profiles and memory/effector phenotypes of HIV-specific and CMV-specific T cells using short-term in vitro stimulation with HIV or CMV peptide pools. Nineteen HIV-positive subjects with progressive disease and twenty healthy, HIV-negative subjects were examined. RESULTS: Among HIV-infected subjects, there were significantly fewer CD8+ IL-2+ T cells responding to HIV compared to CMV, with no significant difference in CD4+ IL-2+ T cells. The majority of CMV-specific T cells in both HIV-negative and HIV-positive subjects appeared to be terminally differentiated effector cells (CD8+ CD27- CD28- CD45RA+ or CD8+ CD27- CD28- CD45RA-). In HIV-positive subjects, the most common phenotype of HIV-specific T cells was intermediate in differentiation (CD8+ CD27+ CD28- CD45RA-). These differences were statistically significant, both as absolute cell frequencies and as percentages. There was a significant correlation between the absolute number of HIV-specific CD8+ IL-2+ T cells and HIV-specific CD8+ CD27- CD28- CD45RA+ terminal effector cells. CONCLUSION: IL-2 production from antigen-specific CD8+ T cells correlates with effector cell differentiation of those cells.

The use of the BD oxygen biosensor system to assess isolated human islets of langerhans: oxygen consumption as a potential measure of islet potency.

The measurement of cellular oxygen consumption rate (OCR) is a potential tool for the assessment of metabolic potency of isolated islets of Langerhans prior to clinical transplantation. We used a commercially available 96-well plate fluoroprobe, the BD Oxygen Biosensor System (OBS), to estimate OCR in 27 human islet preparations, and compared these results to those of concurrent mouse transplantations. OCR was estimated both from the dO2 at steady state and from the transient rate of change of dO2 during the initial culture period immediately after seeding ("dO2 slope"). To demonstrate the validity of the OBS-derived values, it was shown that they scaled linearly with islet equivalent number/DNA concentration and with each other. These measurements were obtained for each preparation of islets incubated in media supplemented with either low (2.2 mM) or high (22 mM) glucose. Concurrently, one to three athymic nude mice were transplanted with 2,000 IEQs under the kidney capsule. The OCR Index, defined as the ratio of the DNA-normalized "dO2 slope" in high glucose to that in low glucose, proved highly predictive of mouse transplant results. Of the 69 mice transplanted, those receiving islets where the OCR Index exceeded 1.27 were 90% likely to reverse within 3 days, whereas those receiving islets with an OCR Index below 1.27 took significantly longer, often failing to reverse at all over a 35-day time period. These results suggest that the OBS could be a useful tool for the pretransplant assessment of islet cell potency.

Identification of compounds that enhance the anti-lymphoma activity of rituximab using flow cytometric high-content screening.

In this report, we describe a new flow cytometry technique termed flow cytometric high-content screening (FC-HCS) which involves semi-automated processing and analysis of multiparameter flow cytometry samples. As a first test of the FC-HCS technique, we used it to screen a 2000-compound library, called the National Cancer Institute (NCI) Diversity Set, to identify agents that would enhance the anti-lymphoma activity of the therapeutic monoclonal antibody rituximab. FC-HCS identified 15 compounds from the Diversity Set that significantly enhanced the ability of rituximab to inhibit cell cycle progression and induce apoptosis in lymphoma cells. The validity of the screening results was confirmed for several compounds using additional assays of cell proliferation, apoptosis and cell growth. The FC-HCS technique was relatively simple and reliable and could process up to 1000 samples/day on a single flow cytometer. The FC-HCS technique may be useful for a variety of applications including drug discovery, immunologic monitoring of patients, functional genomics studies and tissue engineering efforts.

Modeling individual variation in biomarker response to combination immunosuppression with stimulated lymphocyte responses-potential clinical implications.

BACKGROUND: In mitogen-stimulated lymphocyte responses (sLR), cytokines and cell-surface receptors in peripheral human blood lymphocytes (PBL) are sensitive to cyclosporine (CsA), and can predict its in vivo effect with pharmacodynamic (PD) modeling. This is not known for multiple-agent combinations. METHODS: Twenty-five concentration mixtures of CsA (0-1200 ng/ml) plus sirolimus (SRL, 0-30 ng/ml) were added to whole blood from five normal human subjects (NHS) for effect on a limited array of six targets. Effect-concentration relationships were analyzed with E(max) PD equations, and expressed as the range of concentration mixtures associated with one-half of maximal inhibitory effect (EC(50)) on a model biomarker target. This predicted range was examined to see whether it contained representative concentration mixtures of these two drugs, which were associated with a stable post-transplant outcome in a logistic regression model of 1039 clinical trial patients. RESULTS: PD analyses suggested that in NHS samples containing CsA+SRL (n=5), (1) PMA-Ionomycin-stimulated T-cell expression of intracellular IL-2, TNF-alpha, and IFN-gamma was inhibited by CsA, and minimally by SRL, and (2) the two agents inhibited pokeweed mitogen (PWM)-stimulated B-cell expression of CD54 and CD95, but not CD86 (ICAM-1, Fas antigen, and B7.2), synergistically. With CD54 as the model biomarker, contour plots also predicted a wide range of concentration mixtures of the two agents across which an EC(50) could be predicted for CsA+SRL in a population (e.g., CsA-72 ng/ml+SRL 15 ng/ml, n=5), as well as in the individual subject (e.g., CsA-0 ng/ml+SRL-13.75 ng/ml in NHS-D310). Logistic regression analysis of clinical outcomes in 1039 patients suggested that the concentration mixture of CsA congruent with 50-150 ng/ml+SRL congruent with 10 ng/ml was associated with a stable post-transplant course. The EC(50) contour plot for CD54 suggested a nearly identical CsA concentration of 120 ng/ml in the presence of 10 ng/ml of sirolimus. CONCLUSIONS: Our data suggest that pharmacodynamic evaluation of immunosuppressive agents with biomarkers may be an efficient process with which to characterize immunosuppressive effect of combination agents in individual patients and in patient populations.