Spectral Bayesian network theory.

A Bayesian Network (BN) is a probabilistic model that represents a set of variables using a directed acyclic graph (DAG). Current algorithms for learning BN structures from data focus on estimating the edges of a specific DAG, and often lead to many 'likely' network structures. In this paper, we lay the groundwork for an approach that focuses on learning global properties of the DAG rather than exact edges. This is done by defining the structural hypergraph of a BN, which is shown to be related to the inverse-covariance matrix of the network. Spectral bounds are derived for the normalized inverse-covariance matrix, which are shown to be closely related to the maximum indegree of the associated BN.

Impaired Proinflammatory Response in Stringently Defined Otitis-prone Children During Viral Upper Respiratory Infections.

BACKGROUND: Viral upper respiratory infections (URIs) are common and often precipitate acute otitis media (AOM), caused by bacterial otopathogens, in young children. Acute inflammatory responses initiated in the early phase of viral URI contribute to preventing the development of AOM. Stringently-defined otitis-prone (sOP) children are susceptible to recurrent AOM. METHODS: We assessed proinflammatory cytokine and chemokine levels in the nasopharynxes during viral URIs, and examined the different nasopharyngeal responses between viral URI events and the following AOM episodes in both sOP and non-otitis-prone (NOP) children. RESULTS: The sOP children exhibited significantly more AOM episodes per child (8.86-fold higher), viral URIs (P < .0001), and viral URIs followed by AOMs (P < .0001) than the NOP children. The sOP children had lower nasal proinflammatory levels of interleukin (IL)-6 (P = .05), IL-10 (P = .001), tumor necrosis factor (TNF)-α (P = .004), and regulated on activation, normal T-cell-expressed and -secreted (RANTES; P = .002) than NOP children during viral URIs. NOP children had higher levels of IL-6 (P = .02), IL-10 (P = .02), interferon-γ (P = .003), TNF-α (P = .006), IL-1ß (P = .022), monocyte chemoattractant protein 1 (P = .028), RANTES (P = .005), IL-2 (P = .002), and IL-17 (P = .007) during viral URIs versus AOMs following the URIs, when compared to sOP children. CONCLUSIONS: We conclude that sOP children have more frequent viral URIs than NOP children, due to deficient antiviral nasopharyngeal proinflammatory cytokine and chemokine responses.

A model for the regulation of follicular dendritic cells predicts invariant reciprocal-time decay of post-vaccine antibody response.

Follicular dendritic cells (FDC) play a crucial role in the regulation of humoral immunity. They are believed to be responsible for long-term persistence of antibody, due to their role in antibody response induction and their ability to retain antigen in immunogenic form for long periods. In this article, a regulatory control model is proposed which links persistence of humoral immunity with cellular processes associated with FDCs. The argument comprises three elements. The first is a literature review of population-level studies of post-vaccination antibody persistence. It is found that reciprocal-time (∝1/t) decay of antibody levels is widely reported, over a range of ages, observation times and vaccine types. The second element is a mathematical control model for cell population decay for which reciprocal-time decay is a stable attractor. Additionally, control effectors are easily identified, leading to models of homeostatic control of the reciprocal-time decay rate. The final element is a literature review of FDC functionality. This reveals a striking concordance between cell properties required by the model and those widely observed of FDCs, some of which are unique to this cell type. The proposed model is able to unify a wide range of disparate observations of FDC function under one regulatory principle, and to characterize precisely forms of FDC regulation and dysregulation. Many infectious and immunological diseases are increasingly being linked to FDC regulation, therefore a precise understanding of the underlying mechanisms would be of significant benefit for the development of new therapies.

Functional Immune Cell Differences Associated With Low Vaccine Responses in Infants.

BACKGROUND: We sought to understand why some children respond poorly to vaccinations in the first year of life. METHODS: A total of 499 children (6-36 months old) provided serum and peripheral blood mononuclear cell samples after their primary and booster vaccination. Vaccine antigen-specific antibody levels were analyzed with enzyme-linked immunosorbent assay, and frequency of memory B cells, functional T-cell responses, and antigen-presenting cell responses were assessed in peripheral blood mononuclear cell samples with flow cytometric analysis. RESULTS: Eleven percent of children were low vaccine responders, defined a priori as those with subprotective immunoglobulin G antibody levels to ≥66% of vaccines tested. Low vaccine responders generated fewer memory B cells, had reduced activation by CD4(+) and CD8(+) T cells on polyclonal stimulation, and displayed lower major histocompatibility complex II expression by antigen-presenting cells. CONCLUSIONS: We conclude that subprotective vaccine responses in infants are associated with a distinct immunologic profile.

An information theoretic approach to pedigree reconstruction.

Network structure is a dominant feature of many biological systems, both at the cellular level and within natural populations. Advances in genotype and gene expression screening made over the last few decades have permitted the reconstruction of these networks. However, resolution to a single model estimate will generally not be possible, leaving open the question of the appropriate method of formal statistical inference. The nonstandard structure of the problem precludes most traditional statistical methodologies. Alternatively, a Bayesian approach provides a natural methodology for formal inference. Construction of a posterior density on the space of network structures allows formal inference regarding features of network structure using specific marginal posterior distributions. An information theoretic approach to this problem will be described, based on the Minimum Description Length principle. This leads to a Bayesian inference model based on the information content of data rather than on more commonly used probabilistic models. The approach is applied to the problem of pedigree reconstruction based on genotypic data. Using this application, it is shown how the MDL approach is able to provide a truly objective control for model complexity. A two-cohort model is used for a simulation study. The MDL approach is compared to COLONY-2, a well known pedigree reconstruction application. The study highlights the problem of genotyping error modeling. COLONY-2 requires prior error rate estimates, and its accuracy proves to be highly sensitive to these estimates. In contrast, the MDL approach does not require prior error rate estimates, and is able to accurately adjust for genotyping error across the range of models considered.

Effect of oral sodium bicarbonate on fibroblast growth factor-23 in patients with chronic kidney disease: a pilot study.

BACKGROUND: The regulation of fibroblast growth factor-23 (FGF23) secretion in patients with chronic kidney disease (CKD) is incompletely understood. An in vitro study showed that metabolic acidosis increased FGF23 in mouse bone. The objective of this study is to evaluate the effect of oral sodium bicarbonate on circulating FGF23 levels in patients with CKD. METHODS: This was a single-blind pilot study. Twenty adults with estimated glomerular filtration rate between 15-45 mL/min/1.73 m(2) and serum bicarbonate between 20-24 mEq/L were treated with placebo for 2 weeks, followed by increasing doses of oral sodium bicarbonate (0.3, 0.6 and 1.0 mEq/kg/day) in 2 week intervals for a total of 6 weeks. C-terminal FGF23 levels were measured at the initial visit, after 2 weeks of placebo and after 6 weeks of bicarbonate therapy. Wilcoxon matched-pairs signed-rank test was used to compare FGF23 before and after sodium bicarbonate. RESULTS: After 6 weeks of oral sodium bicarbonate, the median FGF23 increased significantly from 150.9 RU/mL (IQR 107.7-267.43) to 191.4 RU/mL (IQR 132.6-316.9) (p = 0.048) and this persisted after excluding participants who received activated vitamin D. CONCLUSIONS: FGF23 increased after short-term oral sodium bicarbonate therapy in patients with CKD and mild metabolic acidosis. It is unclear whether this was due to the alkalinizing effect of sodium bicarbonate or other factors. TRIAL REGISTRATION: The study was registered at ClinicalTrials.gov ( NCT00888290 ) on April 23, 2009.

On non-detects in qPCR data.

MOTIVATION: Quantitative real-time PCR (qPCR) is one of the most widely used methods to measure gene expression. Despite extensive research in qPCR laboratory protocols, normalization and statistical analysis, little attention has been given to qPCR non-detects-those reactions failing to produce a minimum amount of signal. RESULTS: We show that the common methods of handling qPCR non-detects lead to biased inference. Furthermore, we show that non-detects do not represent data missing completely at random and likely represent missing data occurring not at random. We propose a model of the missing data mechanism and develop a method to directly model non-detects as missing data. Finally, we show that our approach results in a sizeable reduction in bias when estimating both absolute and differential gene expression. AVAILABILITY AND IMPLEMENTATION: The proposed algorithm is implemented in the R package, nondetects. This package also contains the raw data for the three example datasets used in this manuscript. The package is freely available at http://mnmccall.com/software and as part of the Bioconductor project.

Comparing 10-yr renal outcomes in deceased donor and living donor liver transplants.

Few studies have explored whether the type of LT, deceased donor LT (DDLT) or living donor LT (LDLT), impacts long-term renal outcomes. We performed a retrospective analysis of 220 LT recipients at our institution to study their renal outcomes at 10 yr. Exclusion criteria were age ≤ 18 yr, graft survival ≤ 6 months, and multiorgan transplants; 108 DDLTs and 62 LDLTs were eligible. At baseline, DDLTs had a lower eGFR than LDLTs and 10.2% of DDLTs were on dialysis as compared to 0% of LDLTs. At 10 yr, seven DDLT and three LDLT recipients required dialysis or renal transplant (p = 0.75). In recipients with graft survival >6 months, DDLTs had a slower decline in eGFR as compared to LDLTs (p < 0.01). Among LDLTs, the decline in eGFR continued over the entire 10-yr period, whereas among DDLTs, the decline in eGFR slowed significantly after six months (p = 0.01). This difference between the two groups was not seen among patients in the highest quartile of baseline eGFR. Patient survival and graft survival were similar. In conclusion, the incidence of end-stage renal disease was similar in both DDLT and LDLT patients, but LDLT recipients seem to have a more sustained decline in eGFR when compared with DDLT recipients.

Affymetrix GeneChip microarray preprocessing for multivariate analyses.

Affymetrix GeneChip microarrays are the most widely used high-throughput technology to measure gene expression, and a wide variety of preprocessing methods have been developed to transform probe intensities reported by a microarray scanner into gene expression estimates. There have been numerous comparisons of these preprocessing methods, focusing on the most common analyses-detection of differential expression and gene or sample clustering. Recently, more complex multivariate analyses, such as gene co-expression, differential co-expression, gene set analysis and network modeling, are becoming more common; however, the same preprocessing methods are typically applied. In this article, we examine the effect of preprocessing methods on some of these multivariate analyses and provide guidance to the user as to which methods are most appropriate.

Statistical evaluation of improvement in RNA secondary structure prediction.

With discovery of diverse roles for RNA, its centrality in cellular functions has become increasingly apparent. A number of algorithms have been developed to predict RNA secondary structure. Their performance has been benchmarked by comparing structure predictions to reference secondary structures. Generally, algorithms are compared against each other and one is selected as best without statistical testing to determine whether the improvement is significant. In this work, it is demonstrated that the prediction accuracies of methods correlate with each other over sets of sequences. One possible reason for this correlation is that many algorithms use the same underlying principles. A set of benchmarks published previously for programs that predict a structure common to three or more sequences is statistically analyzed as an example to show that it can be rigorously evaluated using paired two-sample t-tests. Finally, a pipeline of statistical analyses is proposed to guide the choice of data set size and performance assessment for benchmarks of structure prediction. The pipeline is applied using 5S rRNA sequences as an example.

On the choice of prior density for the Bayesian analysis of pedigree structure.

This article is concerned with the choice of structural prior density for use in a fully Bayesian approach to pedigree inference. It is found that the choice of prior has considerable influence on the accuracy of the estimation. To guide this choice, a scale invariance property is introduced. Under a structural prior with this property, the marginal prior distribution of the local properties of a pedigree node (number of parents, offspring, etc.) does not depend on the number of nodes in the pedigree. Such priors are found to arise naturally by an application of the Minimum Description Length (MDL) principle, under which construction of a prior becomes equivalent to the problem of determining the length of a code required to encode a pedigree, using the principles of information theory. The approach is demonstrated using simulated and actual data, and is compared to two well-known applications, CERVUS and COLONY.

Fitting Boolean networks from steady state perturbation data.

Gene perturbation experiments are commonly used for the reconstruction of gene regulatory networks. Typical experimental methodology imposes persistent changes on the network. The resulting data must therefore be interpreted as a steady state from an altered gene regulatory network, rather than a direct observation of the original network. In this article an implicit modeling methodology is proposed in which the unperturbed network of interest is scored by first modeling the persistent perturbation, then predicting the steady state, which may then be compared to the observed data. This results in a many-to-one inverse problem, so a computational Bayesian approach is used to assess model uncertainty. The methodology is first demonstrated on a number of synthetic networks. It is shown that the Bayesian approach correctly assigns high posterior probability to the network structure and steady state behavior. Further, it is demonstrated that where uncertainty of model features is indicated, the uncertainty may be accurately resolved with further perturbation experiments. The methodology is then applied to the modeling of a gene regulatory network using perturbation data from nine genes which have been shown to respond synergistically to known oncogenic mutations. A hypothetical model emerges which conforms to reported regulatory properties of these genes. Furthermore, the Bayesian methodology is shown to be consistent in the sense that multiple randomized applications of the fitting algorithm converge to an approximately common posterior density on the space of models. Such consistency is generally not feasible for algorithms which report only single models. We conclude that fully Bayesian methods, coupled with models which accurately account for experimental constraints, are a suitable tool for the inference of gene regulatory networks, in terms of accuracy, estimation of model uncertainty, and experimental design.

SIRT3 promotes auditory function in young adult FVB/nJ mice but is dispensable for hearing recovery after noise exposure.

Noise-induced hearing loss (NIHL) affects millions of people worldwide and presents a large social and personal burden. Pharmacological activation of SIRT3, a regulator of the mitochondrial oxidative stress response, has a protective effect on hearing thresholds after traumatic noise damage in mice. In contrast, the role of endogenously activated SIRT3 in hearing recovery has not been established. Here we tested the hypothesis that SIRT3 is required in mice for recovery of auditory thresholds after a noise exposure that confers a temporary threshold shift (TTS). SIRT3-specific immunoreactivity is present in outer hair cells, around the post-synaptic regions of inner hair cells, and faintly within inner hair cells. Prior to noise exposure, homozygous Sirt3-KO mice have slightly but significantly higher thresholds than their wild-type littermates measured by the auditory brainstem response (ABR), but not by distortion product otoacoustic emissions (DPOAE). Moreover, homozygous Sirt3-KO mice display a significant reduction in the progression of their peak 1 amplitude at higher frequencies prior to noise exposure. After exposure to a single sub-traumatic noise dose that does not permanently reduce cochlear function, compromise cell survival, or damage synaptic structures in wild-type mice, there was no difference in hearing function between the two genotypes, measured by ABR and DPOAE. The numbers of hair cells and auditory synapses were similar in both genotypes before and after noise exposure. These loss-of-function studies complement previously published gain-of-function studies and help refine our understanding of SIRT3's role in cochlear homeostasis under different damage paradigms. They suggest that SIRT3 may promote spiral ganglion neuron function. They imply that cellular mechanisms of homeostasis, in addition to the mitochondrial oxidative stress response, act to restore hearing after TTS. Finally, we present a novel application of a biomedical statistical analysis for identifying changes between peak 1 amplitude progressions in ABR waveforms after damage.

Nef gene evolution from a single transmitted strain in acute SIV infection.

BACKGROUND: The acute phase of immunodeficiency virus infection plays a crucial role in determining steady-state virus load and subsequent progression of disease in both humans and nonhuman primates. The acute period is also the time when vaccine-mediated effects on host immunity are likely to exert their major effects on virus infection. Recently we developed a Monte-Carlo (MC) simulation with mathematical analysis of viral evolution during primary HIV-1 infection that enables classification of new HIV-1 infections originating from multiple versus single transmitted viral strains and the estimation of time elapsed following infection. RESULTS: A total of 322 SIV nef SIV sequences, collected during the first 3 weeks following experimental infection of two rhesus macaques with the SIVmac239 clone, were analyzed and found to display a comparable level of genetic diversity, 0.015% to 0.052%, with that of env sequences from acute HIV-1 infection, 0.005% to 0.127%. We confirmed that the acute HIV-1 infection model correctly identified the experimental SIV infections in rhesus macaques as "homogenous" infections, initiated by a single founder strain. The consensus sequence of the sampled strains corresponded to the transmitted sequence as the model predicted. However, measured sequential decrease in diversity at day 7, 11, and 18 post infection violated the model assumption, neutral evolution without any selection. CONCLUSION: While nef gene evolution over the first 3 weeks of SIV infection originating from a single transmitted strain showed a comparable rate of sequence evolution to that observed during acute HIV-1 infection, a purifying selection for the founder nef gene was observed during the early phase of experimental infection of a nonhuman primate.

A note on the calculation of N-statistics.

A class of statistics suitable for testing against equality of multivariate distributions is described by Klebanov and co-workers in 2007. Referred to as N-statistics, their discriminating ability is based on various forms of distance kernels in R(d), the intention being to capture distinct forms of deviation from equality. This makes them particularly suitable for large-scale genomic screening applications, in which such variety of alternatives can be anticipated. One of these kernels, denoted as L(4), introduces weighting by directional densities, hence the evaluation of L(4) requires integration on the unit sphere in R(d). In this note we introduce a methodology for the evaluation of integrals related to L(4). It is shown that for a class of directional densities including, but not limited to, the uniform density L(4) reduces to Euclidean distance. For other cases, the methodology permits a direct interpretation of L(4) in terms of the directional weighting.

Statistical projection of post-vaccination antibody kinetics between dosing schedules.

Determining a recommended dosage schedule is a crucial component of vaccine administration, and is often subject to reassessment. Ideally, recommendations will be supported by multiple arm clinical trials. However, the considerable cost in both resources and time means that a method of predicting post-vaccine humoral antibody levels associated with a hypothetical schedule using data collected from a currently implemented schedule would be of significant benefit to vaccination practice. In this paper we propose such a methodology, which permits statistical estimation of the population mean and standard deviation of log transformed antibody titers of various post-vaccination time points of a hypothetical schedule, using a longitudinal sample of antibody titers from an observed schedule. The method is based on the decomposition of humoral antibody kinetic history into distinct phases, for example, peak phase, decay phase and post-booster phase. The method is feasible because each phase has its own discernable kinetic laws. Of particular interest will be estimation of antibody levels immediately preceding a booster dose (typically the lowest level attained during the schedule), and the antibody levels following a booster dose.

Serum antibody response to Moraxella catarrhalis proteins in stringently defined otitis prone children.

BACKGROUND: Moraxella catarrhalis (Mcat) is a frequent pathogen of acute otitis media (AOM) in young children. Here we prospectively assessed naturally-induced serum antibodies to four Mcat vaccine candidate proteins in stringently defined otitis prone (sOP) and non-otitis prone (NOP) children age 6-36months old following nasopharyngeal (NP) colonization, at onset of AOM and convalescence from AOM. METHODS: Serum IgG and IgM antibody against recombinant Mcat proteins, oligopeptide permease A (OppA), outer membrane protein (OMP) CD, hemagglutinin (Hag), and PilA clade 2 (PilA2), were quantitated by ELISA. RESULTS: During NP colonization by Mcat all four antigens were immunogenic in both sOP and NOP children. However, sOP children had lower antibody responses than NOP children across age 6-36months, similar to our findings for protein vaccine candidates of Streptococcus pneumoniae (Spn) and Nontypeable Haemophilus influenzae (NTHi). sOP children displayed a later and lower peak of antibody rise than NOP children for all four antigens during NP colonization of Mcat. The age-dependent increase of antibody ranked as OppA>Hag5-9>OMP CD>PilA2 in both sOP and NOP children. Lower serum antibody levels to the Mcat antigens were measured in sOP compared to NOP children at the onset of AOM. We did not find a consistent significant increase of antibody at the convalescence phase after an AOM event. CONCLUSIONS: sOP children is a highly vulnerable population that mount lower serum antibody responses to Mcat candidate vaccine proteins compared to NOP children during asymptomatic NP carriage and at onset of AOM.

Home monitoring using wearable radio frequency transmitters.

BACKGROUND: Location tracking of a wearable radio frequency (RF) transmitter in a wireless network is a potentially useful tool for the home monitoring of patients in clinical applications. However, the problem of converting RF signals into accurate estimates of transmitter location remains a significant challenge. OBJECTIVES: We wish to demonstrate that long-term home monitoring using RF transmitters is feasible. Additionally, we conjecture that human motion within familiar environments is confined to relatively small regions of high occupancy. Hence, human motion can be modelled as movement along a network of such high occupancy regions. METHODS AND MATERIALS: Our methodology uses a signal processing technique developed by one of the authors (Almudevar). The technique converts longitudinal RF data into an estimated trajectory which does not depend on explicit location estimates. This approach eliminates the need for a location-signal calibration procedure. Given a long-term trajectory, Gaussian mixture models are used to identify high occupancy regions. The methodology was evaluated using data collected under a study funded by an Everyday Technologies for Alzheimer Care (ETAC) research grant from the Alzheimer's Association. A home monitoring system provided by Home Free Systems was used. RESULTS: The proposed methodology was able to reliably reconstruct trajectories using study data. Regions of high occupancy were identified, and the observed transitions between these regions were found to be spatially and serially coherent. In addition, the trajectory was compared to output from a parallel home sensor network, and a high degree a conformity was evident. CONCLUSION: Long-term home monitoring of human motion is feasible using readily available and easily installable technology. Furthermore, by using suitable signal processing algorithms, the often difficult location-signal calibration process can be bypassed.

Haemophilus influenzae-protein D specific antibody correlate with protection against acute otitis media in young children.

BACKGROUND: Haemophilus influenzae (Hi) causes respiratory infections and pathogenesis of this microbe begins in the human nasopharynx (NP). The objective of this study was to assess the correlation of NP colonization-induced serum antibody levels to Hi protein D with risk of acute otitis media (AOM) in children <2 yr. METHODS: 455 sera from 213 children (age 6-24 months old) were collected when they were colonized with Hi and when the children developed AOM. Presence of Hi during AOM was confirmed by culture of middle ear fluid. Quantitative ELISA was used to determine serum IgG against protein D antigen. RESULTS: Asymptomatic Hi NP colonization reduced the risk of future AOM infections. Higher serum IgG titers against Hi protein D were correlated with reduced future AOM risk. CONCLUSION: Colonization by Hi reduces future AOM risk. Higher antibody levels against protein D correlates with lower risk of AOM caused by Hi.