A reformulation of pLSA for uncertainty estimation and hypothesis testing in bio-imaging.

MOTIVATION: Probabilistic latent semantic analysis (pLSA) is commonly applied to describe mass spectra (MS) images. However, the method does not provide certain outputs necessary for the quantitative scientific interpretation of data. In particular, it lacks assessment of statistical uncertainty and the ability to perform hypothesis testing. We show how linear Poisson modelling advances pLSA, giving covariances on model parameters and supporting χ2 testing for the presence/absence of MS signal components. As an example, this is useful for the identification of pathology in MALDI biological samples. We also show potential wider applicability, beyond MS, using magnetic resonance imaging (MRI) data from colorectal xenograft models. RESULTS: Simulations and MALDI spectra of a stroke-damaged rat brain show MS signals from pathological tissue can be quantified. MRI diffusion data of control and radiotherapy-treated tumours further show high sensitivity hypothesis testing for treatment effects. Successful χ2 and degrees-of-freedom are computed, allowing null-hypothesis thresholding at high levels of confidence. AVAILABILITY AND IMPLEMENTATION: Open-source image analysis software available from TINA Vision, www.tina-vision.net. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Quantifying biological samples using Linear Poisson Independent Component Analysis for MALDI-ToF mass spectra.

Motivation: Matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI) facilitates the analysis of large organic molecules. However, the complexity of biological samples and MALDI data acquisition leads to high levels of variation, making reliable quantification of samples difficult. We present a new analysis approach that we believe is well-suited to the properties of MALDI mass spectra, based upon an Independent Component Analysis derived for Poisson sampled data. Simple analyses have been limited to studying small numbers of mass peaks, via peak ratios, which is known to be inefficient. Conventional PCA and ICA methods have also been applied, which extract correlations between any number of peaks, but we argue makes inappropriate assumptions regarding data noise, i.e. uniform and Gaussian. Results: We provide evidence that the Gaussian assumption is incorrect, motivating the need for our Poisson approach. The method is demonstrated by making proportion measurements from lipid-rich binary mixtures of lamb brain and liver, and also goat and cow milk. These allow our measurements and error predictions to be compared to ground truth. Availability and implementation: Software is available via the open source image analysis system TINA Vision, www.tina-vision.net. Contact: paul.tar@manchester.ac.uk. Supplementary information: Supplementary data are available at Bioinformatics online.

Inactivated influenza vaccine (IIV) in children <2 years of age: examination of selected adverse events reported to the Vaccine Adverse Event Reporting System (VAERS) after thimerosal-free or thimerosal-containing vaccine.

Thimerosal as a preservative (in all but trace amounts) was removed from vaccines used in infants starting in the late 1990s, though the preservative-including inactivated influenza vaccine is still available for use in individuals >or=6 months of age. We compared the proportion of injection site reactions, rash, and infections reported to the Vaccine Adverse Event Reporting System (VAERS) after preservative-free (PFV), preservative-including (PIV), and preservative unknown (PUV) vaccines in reports from 7/1/2004 to 1/4/2006. There were 145, 175, and 216 reports after vaccination with PFV, PIV, and PUV, respectively. The most frequently reported coding terms (fever, rash, and urticaria) were seen in similar proportions in the PFV, PIV, and PUV groups. No difference was detected in the proportion of injection site reactions (ISR), rash, or infections in the PIV, PFV, and PUV reports. Keeping in mind the inherent limitations of VAERS, including underreporting and potential reporting biases, we conclude that there were no substantial differences in the proportion of rash, ISR, and infection reports in the PIV, PFV and PUV reports in infants.

Comparison of military and civilian reporting rates for smallpox vaccine adverse events.

INTRODUCTION: US smallpox vaccination (SMA) started most recently in December 2002. Military and civilian personnel report adverse events (AEs) to the Vaccine Adverse Event Reporting System (VAERS), a surveillance system that relies on spontaneous reports. Although reported rates of probable myo/pericarditis after SMA in the literature are similar between military personnel and civilian healthcare workers, some civilian AE reporting rates after SMA appeared higher than those in the military. OBJECTIVE: Determine if SMA-associated reporting rates are different in civilians than in the military, considering age, sex, seriousness, and expectedness of the AE, as well as self-reporting. METHODS: Numerators were SMA reports in VAERS from 12/12/02 to 3/1/04. Limitations of VAERS include underreporting and lack of diagnostic confirmation. Denominators were number of military and civilian vaccinees. RESULTS: Reporting rates stratified by age and sex of serious and non-serious AEs were significantly higher in civilian than military personnel ages <55 years (rate ratios 4-27). These rate ratios decreased with increasing age. CONCLUSIONS: Reporting rates in VAERS differed significantly and substantially in civilians compared to military personnel <55 years of age. Differences in stimulated passive surveillance systems, and AE reporting practices, including the 'threshold' for reporting most likely explain these findings. These results suggest that in the case of smallpox vaccine AEs, there may be systematic differences in reporting completeness between the civilian and military sectors, and that passive surveillance data should be interpreted with caution.

Activated lymphocytes promote endothelial cell detachment from matrix: a role for modulation of endothelial cell beta 1 integrin affinity.

In vivo, MHC class I-restricted injury of allogeneic tissue or cells infected by intracellular pathogens occurs in the absence of classical cytolytic effector mechanisms and Ab. Modulation of the target cell adhesion to matrix may be an additional mechanism used to injure vascular or epithelial cells in inflammation. We studied the mechanisms of human umbilical vein endothelial cell (EC) detachment from matrix-coated plastic following contact by concanamycin A-treated lymphocytes as an in vitro model of perforin-independent modulation of EC basement membrane adhesion. Human PBL were depleted of monocytes, stimulated, then added to an EC monolayer plated on either fibronectin or type I collagen matrices. Activated, but not resting, PBL induced progressive EC detachment from the underlying matrix. Injury of the EC monolayer required direct cell contact with the activated lymphocytes because no detachment was seen when the PBL were placed above a Transwell membrane. Moreover plasma membranes prepared from activated but not resting PBL induced EC detachment. Adherent EC stimulated with activated PBL did not show evidence of apoptosis using TUNEL and annexin V staining at time points before EC detachment was observed. Finally, neither the matrix metalloproteinase inhibitors o-phenanthroline and BB-94 nor aprotinin blocked EC detachment. However, activation of EC beta1 integrin using mAb TS2/16 or Mg2+ decreased EC detachment. These data indicate that cell-cell contact between activated PBL and EC reduces adhesion of EC to the underlying matrix, at least in part by inducing changes in the affinity of the endothelial beta 1 integrin.