Background: Electronic medical record (EMR) computed algorithms allow investigators to screen thousands of patient records to identify specific disease cases. No computed algorithms have been developed to detect all cases of human immunodeficiency virus (HIV) infection using administrative, laboratory, and clinical documentation data outside of the Veterans Health Administration. We developed novel EMR-based algorithms for HIV detection and validated them in a cohort of subjects in the Duke University Health System (DUHS). Methods: We created 2 novel algorithms to identify HIV-infected subjects. Algorithm 1 used laboratory studies and medications to identify HIV-infected subjects, whereas Algorithm 2 used International Classification of Diseases, Ninth Revision (ICD-9) codes, medications, and laboratory testing. We applied the algorithms to a well-characterized cohort of patients and validated both against the gold standard of physician chart review. We determined sensitivity, specificity, and prevalence of HIV between 2007 and 2011 in patients seen at DUHS. Results: A total of 172 271 patients were detected with complete data; 1063 patients met algorithm criteria for HIV infection. In all, 970 individuals were identified by both algorithms, 78 by Algorithm 1 alone, and 15 by Algorithm 2 alone. The sensitivity and specificity of each algorithm were 78% and 99%, respectively, for Algorithm 1 and 77% and 100% for Algorithm 2. The estimated prevalence of HIV infection at DUHS between 2007 and 2011 was 0.6%. Conclusions: EMR-based phenotypes of HIV infection are capable of detecting cases of HIV-infected adults with good sensitivity and specificity. These algorithms have the potential to be adapted to other EMR systems, allowing for the creation of cohorts of patients across EMR systems.

OBJECTIVE: We assessed the sensitivity and specificity of 8 electronic health record (EHR)-based phenotypes for diabetes mellitus against gold-standard American Diabetes Association (ADA) diagnostic criteria via chart review by clinical experts. MATERIALS AND METHODS: We identified EHR-based diabetes phenotype definitions that were developed for various purposes by a variety of users, including academic medical centers, Medicare, the New York City Health Department, and pharmacy benefit managers. We applied these definitions to a sample of 173 503 patients with records in the Duke Health System Enterprise Data Warehouse and at least 1 visit over a 5-year period (2007-2011). Of these patients, 22 679 (13%) met the criteria of 1 or more of the selected diabetes phenotype definitions. A statistically balanced sample of these patients was selected for chart review by clinical experts to determine the presence or absence of type 2 diabetes in the sample. RESULTS: The sensitivity (62-94%) and specificity (95-99%) of EHR-based type 2 diabetes phenotypes (compared with the gold standard ADA criteria via chart review) varied depending on the component criteria and timing of observations and measurements. DISCUSSION AND CONCLUSIONS: Researchers using EHR-based phenotype definitions should clearly specify the characteristics that comprise the definition, variations of ADA criteria, and how different phenotype definitions and components impact the patient populations retrieved and the intended application. Careful attention to phenotype definitions is critical if the promise of leveraging EHR data to improve individual and population health is to be fulfilled.

Electronic Health Record (EHR) phenotyping utilizes patient data captured through normal medical practice, to identify features that may represent computational medical phenotypes. These features may be used to identify at-risk patients and improve prediction of patient morbidity and mortality. We present a novel deep multi-modality architecture for EHR analysis (applicable to joint analysis of multiple forms of EHR data), based on Poisson Factor Analysis (PFA) modules. Each modality, composed of observed counts, is represented as a Poisson distribution, parameterized in terms of hidden binary units. Information from different modalities is shared via a deep hierarchy of common hidden units. Activation of these binary units occurs with probability characterized as Bernoulli- Poisson link functions, instead of more traditional logistic link functions. In addition, we demonstrate that PFA modules can be adapted to discriminative modalities. To compute model parameters, we derive efficient Markov Chain Monte Carlo (MCMC) inference that scales efficiently, with significant computational gains when compared to related models based on logistic link functions. To explore the utility of these models, we apply them to a subset of patients from the Duke-Durham patient cohort. We identified a cohort of over 16,000 patients with Type 2 Diabetes Mellitus (T2DM) based on diagnosis codes and laboratory tests out of our patient population of over 240,000. Examining the common hidden units uniting the PFA modules, we identify patient features that represent medical concepts. Experiments indicate that our learned features are better able to predict mortality and morbidity than clinical features identified previously in a large-scale clinical trial.

Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records.

SUMMARY: A non-parametric Bayesian factor model is proposed for joint analysis of multi-platform genomics data. The approach is based on factorizing the latent space (feature space) into a shared component and a data-specific component with the dimensionality of these components (spaces) inferred via a beta-Bernoulli process. The proposed approach is demonstrated by jointly analyzing gene expression/copy number variations and gene expression/methylation data for ovarian cancer patients, showing that the proposed model can potentially uncover key drivers related to cancer. AVAILABILITY AND IMPLEMENTATION: The source code for this model is written in MATLAB and has been made publicly available at https://sites.google.com/site/jointgenomics/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: The goal of many proteomics experiments is to determine the abundance of proteins in biological samples, and the variation thereof in various physiological conditions. High-throughput quantitative proteomics, specifically label-free LC-MS/MS, allows rapid measurement of thousands of proteins, enabling large-scale studies of various biological systems. Prior to analyzing these information-rich datasets, raw data must undergo several computational processing steps. We present a method to address one of the essential steps in proteomics data processing--the matching of peptide measurements across samples. RESULTS: We describe a novel method for label-free proteomics data alignment with the ability to incorporate previously unused aspects of the data, particularly ion mobility drift times and product ion information. We compare the results of our alignment method to PEPPeR and OpenMS, and compare alignment accuracy achieved by different versions of our method utilizing various data characteristics. Our method results in increased match recall rates and similar or improved mismatch rates compared to PEPPeR and OpenMS feature-based alignment. We also show that the inclusion of drift time and product ion information results in higher recall rates and more confident matches, without increases in error rates. CONCLUSIONS: Based on the results presented here, we argue that the incorporation of ion mobility drift time and product ion information are worthy pursuits. Alignment methods should be flexible enough to utilize all available data, particularly with recent advancements in experimental separation methods.

BACKGROUND: Label-free quantitative proteomics holds a great deal of promise for the future study of both medicine and biology. However, the data generated is extremely intricate in its correlation structure, and its proper analysis is complex. There are issues with missing identifications. There are high levels of correlation between many, but not all, of the peptides derived from the same protein. Additionally, there may be systematic shifts in the sensitivity of the machine between experiments or even through time within the duration of a single experiment. RESULTS: We describe a hierarchical model for analyzing unbiased, label-free proteomics data which utilizes the covariance of peptide expression across samples as well as MS/MS-based identifications to group peptides-a strategy we call metaprotein expression modeling. Our metaprotein model acknowledges the possibility of misidentifications, post-translational modifications and systematic differences between samples due to changes in instrument sensitivity or differences in total protein concentration. In addition, our approach allows us to validate findings from unbiased, label-free proteomics experiments with further unbiased, label-free proteomics experiments. Finally, we demonstrate the clinical/translational utility of the model for building predictors capable of differentiating biological phenotypes as well as for validating those findings in the context of three novel cohorts of patients with Hepatitis C. CONCLUSIONS: Mass-spectrometry proteomics is quickly becoming a powerful tool for studying biological and translational questions. Making use of all of the information contained in a particular set of data will be critical to the success of those endeavors. Our proposed model represents an advance in the ability of statistical models of proteomic data to identify and utilize correlation between features. This allows validation of predictors without translation to targeted assays in addition to informing the choice of targets when it is appropriate to generate those assays.

BACKGROUND: The biological phenotype of a cell, such as a characteristic visual image or behavior, reflects activities derived from the expression of collections of genes. As such, an ability to measure the expression of these genes provides an opportunity to develop more precise and varied sets of phenotypes. However, to use this approach requires computational methods that are difficult to implement and apply, and thus there is a critical need for intelligent software tools that can reduce the technical burden of the analysis. Tools for gene expression analyses are unusually difficult to implement in a user-friendly way because their application requires a combination of biological data curation, statistical computational methods, and database expertise. RESULTS: We have developed SIGNATURE, a web-based resource that simplifies gene expression signature analysis by providing software, data, and protocols to perform the analysis successfully. This resource uses bayesian methods for processing gene expression data coupled with a curated database of gene expression signatures, all carried out within a GenePattern web interface for easy use and access. CONCLUSIONS: SIGNATURE is available for public use at http://genepattern.genome.duke.edu/signature/.

BACKGROUND: Nonparametric Bayesian techniques have been developed recently to extend the sophistication of factor models, allowing one to infer the number of appropriate factors from the observed data. We consider such techniques for sparse factor analysis, with application to gene-expression data from three virus challenge studies. Particular attention is placed on employing the Beta Process (BP), the Indian Buffet Process (IBP), and related sparseness-promoting techniques to infer a proper number of factors. The posterior density function on the model parameters is computed using Gibbs sampling and variational Bayesian (VB) analysis. RESULTS: Time-evolving gene-expression data are considered for respiratory syncytial virus (RSV), Rhino virus, and influenza, using blood samples from healthy human subjects. These data were acquired in three challenge studies, each executed after receiving institutional review board (IRB) approval from Duke University. Comparisons are made between several alternative means of per-forming nonparametric factor analysis on these data, with comparisons as well to sparse-PCA and Penalized Matrix Decomposition (PMD), closely related non-Bayesian approaches. CONCLUSIONS: Applying the Beta Process to the factor scores, or to the singular values of a pseudo-SVD construction, the proposed algorithms infer the number of factors in gene-expression data. For real data the "true" number of factors is unknown; in our simulations we consider a range of noise variances, and the proposed Bayesian models inferred the number of factors accurately relative to other methods in the literature, such as sparse-PCA and PMD. We have also identified a "pan-viral" factor of importance for each of the three viruses considered in this study. We have identified a set of genes associated with this pan-viral factor, of interest for early detection of such viruses based upon the host response, as quantified via gene-expression data.

Tumor microenvironmental stresses, such as hypoxia and lactic acidosis, play important roles in tumor progression. Although gene signatures reflecting the influence of these stresses are powerful approaches to link expression with phenotypes, they do not fully reflect the complexity of human cancers. Here, we describe the use of latent factor models to further dissect the stress gene signatures in a breast cancer expression dataset. The genes in these latent factors are coordinately expressed in tumors and depict distinct, interacting components of the biological processes. The genes in several latent factors are highly enriched in chromosomal locations. When these factors are analyzed in independent datasets with gene expression and array CGH data, the expression values of these factors are highly correlated with copy number alterations (CNAs) of the corresponding BAC clones in both the cell lines and tumors. Therefore, variation in the expression of these pathway-associated factors is at least partially caused by variation in gene dosage and CNAs among breast cancers. We have also found the expression of two latent factors without any chromosomal enrichment is highly associated with 12q CNA, likely an instance of "trans"-variations in which CNA leads to the variations in gene expression outside of the CNA region. In addition, we have found that factor 26 (1q CNA) is negatively correlated with HIF-1alpha protein and hypoxia pathways in breast tumors and cell lines. This agrees with, and for the first time links, known good prognosis associated with both a low hypoxia signature and the presence of CNA in this region. Taken together, these results suggest the possibility that tumor segmental aneuploidy makes significant contributions to variation in the lactic acidosis/hypoxia gene signatures in human cancers and demonstrate that latent factor analysis is a powerful means to uncover such a linkage.