Determination of sex-specific 99th percentile upper reference limits for a point of care high sensitivity cardiac troponin I assay.

OBJECTIVES: High sensitivity (hs) cardiac troponin (cTn) assays are defined per the IFCC Committee on Clinical Application of Cardiac Biomarker (C-CB) by the ability to measure ≥ 50% of concentrations greater than the limit of detection (LoD) with an impression of ≤10% at sex-specific 99th percentiles. Our study determined the sex-specific 99th percentile upper reference limits for males and females utilizing heparinized plasma from AACC universal sample bank for the Siemens point of care (POC) Atellica® VTLi hs-cTnI immunoassay. METHODS: Apparently healthy subjects, included overall 693, males 363, and females 330, following exclusionary surrogate biomarker use of hemoglobin A1c, NT-proBNP, and eGFR, along with statin medication. hs-cTnI was measured in a central laboratory, on multiple POC Atellica® VTLi immunoassay analyzers. The LoD was 1.24 ng/L and the 10%CV concentration was 6.7 ng/L. 99th percentile URLs were determined by the nonparametric (NP) method. RESULTS: Histograms of the hs-cTnI concentrations (ng/L) for males and females were used to visualize the distributions and concentrations in men and women and differed significantly (pre- and post-exclusion, both p <0.001). 99th percentile URLs were: overall 23 ng/L (90% CI 20-32 ng/L); male 27 ng/L (CI 21-37 ng/L); female 18 ng/L (CI 9-78 ng/L). The percentages of subjects having a measurable concentration ≥ the LoD were: overall 83.7%, male 87.3%, female 79.7%. CONCLUSIONS: Our findings show the novel POC Atellica® VTLi hs-cTnI assay meets the designation of a 'high-sensitivity' assay using heparinized plasma.

Optimization of the empirical antibiotic choice during the treatment of acute prosthetic joint infections: a retrospective analysis of 91 patients.

Background and purpose - The preferred treatment of an acute prosthetic joint infection (PJI) is debridement, antibiotics, irrigation and retention of the prosthesis (DAIR). The antibiotic treatment consists of an empirical and targeted phase. In the empirical phase, intravenous antibiotics are started after surgery before micro-organisms are determined in microbiological cultures. Which empirical antibiotic is used differs between hospitals, partly reflecting geographic differences in susceptibility spectrums. We investigated whether flucloxacillin should remain the antibiotic of choice in our hospital for empiric treatment of acute PJI with DAIR. Patients and methods - We retrospectively analyzed 91 patients treated for PJI with DAIR between 2012 and 2016. The susceptibility of micro-organisms was determined in multiple cultures of periprosthetic tissue and synovial fluid for 3 antibiotics: amoxicillin/clavulanic acid, cefazolin, and flucloxacillin. Results - Positive microbiological cultures from 68 patients were analyzed. Staphylococcus aureus was the predominant pathogen, cultured in half of the patients. In one-third of patients more than 1 micro-organism was found. On a patient level, the data showed that 65% were responsive to flucloxacillin, 76% to amoxicillin/clavulanic acid, and 79% to cefazolin. Interpretation - Flucloxacillin appeared to be a suboptimal choice in our patient population treated with DAIR. We therefore changed our practice to cefazolin as the preferred antibiotic in the empirical treatment of acute PJI with DAIR.

Influenza Virus Reassortment Is Enhanced by Semi-infectious Particles but Can Be Suppressed by Defective Interfering Particles.

A high particle to infectivity ratio is a feature common to many RNA viruses, with ~90-99% of particles unable to initiate a productive infection under low multiplicity conditions. A recent publication by Brooke et al. revealed that, for influenza A virus (IAV), a proportion of these seemingly non-infectious particles are in fact semi-infectious. Semi-infectious (SI) particles deliver an incomplete set of viral genes to the cell, and therefore cannot support a full cycle of replication unless complemented through co-infection. In addition to SI particles, IAV populations often contain defective-interfering (DI) particles, which actively interfere with production of infectious progeny. With the aim of understanding the significance to viral evolution of these incomplete particles, we tested the hypothesis that SI and DI particles promote diversification through reassortment. Our approach combined computational simulations with experimental determination of infection, co-infection and reassortment levels following co-inoculation of cultured cells with two distinct influenza A/Panama/2007/99 (H3N2)-based viruses. Computational results predicted enhanced reassortment at a given % infection or multiplicity of infection with increasing semi-infectious particle content. Comparison of experimental data to the model indicated that the likelihood that a given segment is missing varies among the segments and that most particles fail to deliver ≥1 segment. To verify the prediction that SI particles augment reassortment, we performed co-infections using viruses exposed to low dose UV. As expected, the introduction of semi-infectious particles with UV-induced lesions enhanced reassortment. In contrast to SI particles, inclusion of DI particles in modeled virus populations could not account for observed reassortment outcomes. DI particles were furthermore found experimentally to suppress detectable reassortment, relative to that seen with standard virus stocks, most likely by interfering with production of infectious progeny from co-infected cells. These data indicate that semi-infectious particles increase the rate of reassortment and may therefore accelerate adaptive evolution of IAV.

Antigenic Maps of Influenza A(H3N2) Produced With Human Antisera Obtained After Primary Infection.

BACKGROUND: Antigenic characterization of influenza viruses is typically based on hemagglutination inhibition (HI) assay data for viral isolates tested against strain-specific postinfection ferret antisera. Here, similar virus characterizations were performed using serological data from humans with primary influenza A(H3N2) infection. METHODS: We screened sera collected between 1995 and 2011 from children between 9 and 24 months of age for influenza virus antibodies, performed HI tests for the positive sera against 23 influenza viruses isolated between 1989 and 2011, and measured HI titers of antisera against influenza A(H3N2) from 24 ferrets against the same panel of viruses. RESULTS: Of the 17 positive human sera, 6 had a high response, showing HI patterns that would be expected from primary infection antisera, while 11 sera had lower, more dispersed patterns of reactivity that are not easily explained. The antigenic map based on the high-response human HI data was similar to the map created using ferret data. CONCLUSIONS: Although the overall structure of the ferret and human antigenic maps is similar, local differences in virus positions indicate that the human and ferret immune system might see antigenic properties of viruses differently. Further studies are needed to establish the degree of similarity between serological patterns in ferret and human data.

Expected Effect of Deleterious Mutations on Within-Host Adaptation of Pathogens.

UNLABELLED: Adaptation is a common theme in both pathogen emergence, for example, in zoonotic cross-species transmission, and pathogen control, where adaptation might limit the effect of the immune response and antiviral treatment. When such evolution requires deleterious intermediate mutations, fitness ridges and valleys arise in the pathogen's fitness landscape. The effect of deleterious intermediate mutations on within-host pathogen adaptation is examined with deterministic calculations, appropriate for pathogens replicating in large populations with high error rates. The effect of deleterious intermediate mutations on pathogen adaptation is smaller than their name might suggest: when two mutations are required and each individual single mutation is fully deleterious, the pathogen can jump across the fitness valley by obtaining two mutations at once, leading to a proportion of adapted mutants that is 20-fold lower than that in the situation where the fitness of all mutants is neutral. The negative effects of deleterious intermediates are typically substantially smaller and outweighed by the fitness advantages of the adapted mutant. Moreover, requiring a specific mutation order has a substantially smaller effect on pathogen adaptation than the effect of all intermediates being deleterious. These results can be rationalized when the number of routes of mutation available to the pathogen is calculated, providing a simple approach to estimate the effect of deleterious mutations. The calculations discussed here are applicable when the effect of deleterious mutations on the within-host adaptation of pathogens is assessed, for example, in the context of zoonotic emergence, antigenic escape, and drug resistance. IMPORTANCE: Adaptation is critical for pathogens after zoonotic transmission into a new host species or to achieve antigenic immune escape and drug resistance. Using a deterministic approach, the effects of deleterious intermediate mutations on pathogen adaptation were calculated while avoiding commonly made simplifications that do not apply to large pathogen populations replicating with high mutation rates. Perhaps unexpectedly, pathogen adaptation does not halt when the intermediate mutations are fully deleterious. The negative effects of deleterious mutations are substantially outweighed by the fitness gains of adaptation. To gain an understanding of the effect of deleterious mutations on pathogen adaptation, a simple approach that counts the number of routes available to the pathogen with and without deleterious intermediate mutations is introduced. This methodology enables a straightforward calculation of the proportion of the pathogen population that will cross a fitness valley or traverse a fitness ridge, without reverting to more complicated mathematical models.

Mapping Enterovirus A71 Antigenic Determinants from Viral Evolution.

UNLABELLED: Human enterovirus A71 (EV-A71) belongs to the Enterovirus A species in the Picornaviridae family. Several vaccines against EV-A71, a disease causing severe neurological complications or even death, are currently under development and being tested in clinical trials, and preventative vaccination programs are expected to start soon. To characterize the potential for antigenic change of EV-A71, we compared the sequences of two antigenically diverse genotype B4 and B5 strains of EV-A71 and identified substitutions at residues 98, 145, and 164 in the VP1 capsid protein as antigenic determinants. To examine the effects of these three substitutions on antigenicity, we constructed a series of recombinant viruses containing different mutation combinations at these three residues with a reverse genetics system and then investigated the molecular basis of antigenic changes with antigenic cartography. We found that a novel EV-A71 mutant, containing lysine, glutamine, and glutamic acid at the respective residues 98, 145, and 164 in the VP1 capsid protein, exhibited neutralization reduction against patients' antisera and substantially increased virus binding ability to human cells. These observations indicated that this low-neutralization-reactive EV-A71 VP1-98K/145Q/164E mutant potentially increases viral binding ability and that surveillance studies should look out for these mutants, which could compromise vaccine efficacy. IMPORTANCE: Emerging and reemerging EV-A71 viruses can cause severe neurological etiology, primarily affecting children, especially around Asia-Pacific countries. We identified a set of mutations in EV-A71 that both reduced neutralization activity against humoral immunity in antisera of patients and healthy adults and greatly increased the viral binding ability to cells. These findings provide important insights for EV-A71 antigenic determinants and emphasize the importance of continuous surveillance, especially after EV-A71 vaccination programs begin.

Hyperspectral visualization of mass spectrometry imaging data.

The acquisition of localized molecular spectra with mass spectrometry imaging (MSI) has a great, but as yet not fully realized, potential for biomedical diagnostics and research. The methodology generates a series of mass spectra from discrete sample locations, which is often analyzed by visually interpreting specifically selected images of individual masses. We developed an intuitive color-coding scheme based on hyperspectral imaging methods to generate a single overview image of this complex data set. The image color-coding is based on spectral characteristics, such that pixels with similar molecular profiles are displayed with similar colors. This visualization strategy was applied to results of principal component analysis, self-organizing maps and t-distributed stochastic neighbor embedding. Our approach for MSI data analysis, combining automated data processing, modeling and display, is user-friendly and allows both the spatial and molecular information to be visualized intuitively and effectively.

Statistical total correlation spectroscopy scaling for enhancement of metabolic information recovery in biological NMR spectra.

The high level of complexity in nuclear magnetic resonance (NMR) metabolic spectroscopic data sets has fueled the development of experimental and mathematical techniques that enhance latent biomarker recovery and improve model interpretability. We previously showed that statistical total correlation spectroscopy (STOCSY) can be used to edit NMR spectra to remove drug metabolite signatures that obscure metabolic variation of diagnostic interest. Here, we extend this "STOCSY editing" concept to a generalized scaling procedure for NMR data that enhances recovery of latent biochemical information and improves biological classification and interpretation. We call this new procedure STOCSY-scaling (STOCSY(S)). STOCSY(S) exploits the fixed proportionality in a set of NMR spectra between resonances from the same molecule to suppress or enhance features correlated with a resonance of interest. We demonstrate this new approach using two exemplar data sets: (a) a streptozotocin rat model (n = 30) of type 1 diabetes and (b) a human epidemiological study utilizing plasma NMR spectra of patients with metabolic syndrome (n = 67). In both cases significant biomarker discovery improvement was observed by using STOCSY(S): the approach successfully suppressed interfering NMR signals from glucose and lactate that otherwise dominate the variation in the streptozotocin study, which then allowed recovery of biomarkers such as glycine, which were otherwise obscured. In the metabolic syndrome study, we used STOCSY(S) to enhance variation from the high-density lipoprotein cholesterol peak, improving the prediction of individuals with metabolic syndrome from controls in orthogonal projections to latent structures discriminant analysis models and facilitating the biological interpretation of the results. Thus, STOCSY(S) is a versatile technique that is applicable in any situation in which variation, either biological or otherwise, dominates a data set at the expense of more interesting or important features. This approach is generally appropriate for many types of NMR-based complex mixture analyses and hence for wider applications in bioanalytical science.

Robust data processing and normalization strategy for MALDI mass spectrometric imaging.

Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) provides localized information about the molecular content of a tissue sample. To derive reliable conclusions from MSI data, it is necessary to implement appropriate processing steps in order to compare peak intensities across the different pixels comprising the image. Here, we review commonly used normalization methods, and propose a rational data processing strategy, for robust evaluation and modeling of MSI data. The approach includes newly developed heuristic methods for selecting biologically relevant peaks and pixels to reduce the size of a data set and remove the influence of the applied MALDI matrix. The methods are demonstrated on a MALDI MSI data set of a sagittal section of rat brain (4750 bins, m/z = 50-1000, 111 × 185 pixels) and the proposed preferred normalization method uses the median intensity of selected peaks, which were determined to be independent of the MALDI matrix. This was found to effectively compensate for a range of known limitations associated with the MALDI process and irregularities in MS image sampling routines. This new approach is relevant for processing of all MALDI MSI data sets, and thus likely to have impact in biomarker profiling, preclinical drug distribution studies, and studies addressing underlying molecular mechanisms of tissue pathology.

Chemical shifts in nucleic acids studied by density functional theory calculations and comparison with experiment.

NMR chemical shifts are highly sensitive probes of local molecular conformation and environment and form an important source of structural information. In this study, the relationship between the NMR chemical shifts of nucleic acids and the glycosidic torsion angle, χ, has been investigated for the two commonly occurring sugar conformations. We have calculated by means of DFT the chemical shifts of all atoms in the eight DNA and RNA mono-nucleosides as a function of these two variables. From the DFT calculations, structures and potential energy surfaces were determined by using constrained geometry optimizations at the BP86/TZ2P level of theory. The NMR parameters were subsequently calculated by single-point calculations at the SAOP/TZ2P level of theory. Comparison of the (1)H and (13)C NMR shifts calculated for the mono-nucleosides with the shifts determined by NMR spectroscopy for nucleic acids demonstrates that the theoretical shifts are valuable for the characterization of nucleic acid conformation. For example, a clear distinction can be made between χ angles in the anti and syn domains. Furthermore, a quantitative determination of the χ angle in the syn domain is possible, in particular when (13)C and (1)H chemical shift data are combined. The approximate linear dependence of the C1' shift on the χ angle in the anti domain provides a good estimate of the angle in this region. It is also possible to derive the sugar conformation from the chemical shift information. The DFT calculations reported herein were performed on mono-nucleosides, but examples are also provided to estimate intramolecularly induced shifts as a result of hydrogen bonding, polarization effects, or ring-current effects.

Point-of-Care: Roadmap for Analytical Characterization and Validation of a High-Sensitivity Cardiac Troponin I Assay in Plasma and Whole Blood Matrices.

BACKGROUND: High-sensitivity cardiac troponin (hs-cTn) assays enable more precise use of traditional diagnostic strategies and earlier rule-out/rule-in at 0/1 h or 0/2 h after presentation of acute myocardial infarction (AMI). Availability of hs-cTn measurements at point-of-care (POC) can improve timely management of AMI patients. A roadmap for regulatory and analytical validation is exemplified with studies with the Atellica VTLi hs-cTnI at POC. METHODS: High-sensitivity performance was assessed with AACC/IFCC expert recommendations. Clinical Laboratory Standards Institute protocols were used for characterizing limit of blank, limit of detection (LoD), limit of quantitation (LoQ), 10% CV, precision, linearity, and analytic specificity with several reagent lots. Bland-Altman, Passing-Bablok, and hematocrit bias plots compared hs-cTnI measurement in lithium-heparin plasma (PL) and whole blood (WB) matrices. RESULTS: LoB was 0.55 ng/L; LoD and LoQ were 1.24 ng/L and 2.1 ng/Lm for PL and 1.60 ng/L and 3.7 ng/L for WB, respectively. The male 99th percentile is 27 ng/L, and female 99th percentile upper reference limit is 18 ng/L; 10% CVs were 6.7 ng/L for PL and 8.9 ng/L for WB. Also ≥50% of hs-cTnI values for healthy cohorts exceeded the LoD, confirming high-sensitivity performance. Linearity spanned from LoQ to 1250 ng/L. Specificity was >90% for 40 potential interferences; no hook effect was detected. WB and PL correlation was WB = 1.02*plasma + 0.3 ng/L (r = 0.996, n = 152). No hs-cTnI association with hematocrit was detected (R2 = 0.003). CONCLUSION: This analytical roadmap showed high-sensitivity performance, good analytic characteristics, and excellent PL and WB agreement for the Atellica VTLi hs-cTnI POC system. Essential clinical performance studies in patients by intended POC users may now commence.

Systems parasitology: effects of Fasciola hepatica on the neurochemical profile in the rat brain.

We characterize the integrated response of a rat host to the liver fluke Fasciola hepatica using a combination of (1)H nuclear magnetic resonance spectroscopic profiles (liver, kidney, intestine, brain, spleen, plasma, urine, feces) and multiplex cytokine markers of systemic inflammation. Multivariate mathematical models were built to describe the main features of the infection at the systems level. In addition to the expected modulation of hepatic choline and energy metabolism, we found significant perturbations of the nucleotide balance in the brain, together with increased plasma IL-13, suggesting a shift toward modulation of immune reactions to minimize inflammatory damage, which may favor the co-existence of the parasite in the host. Subsequent analysis of brain extracts from other trematode infection models (i.e. Schistosoma mansoni, and Echinostoma caproni) did not elicit a change in neural nucleotide levels, indicating that the neural effects of F. hepatica infection are specific. We propose that the topographically extended response to invasion of the host as characterized by the modulated global metabolic phenotype is stratified across several bio-organizational levels and reflects the direct manipulation of host-nucleotide balance.

Evaluation of full-resolution J-resolved 1H NMR projections of biofluids for metabonomics information retrieval and biomarker identification.

Spectroscopic profiling of biological samples is an integral part of metabolically driven top-down systems biology and can be used for identifying biomarkers of toxicity and disease. However, optimal biomarker information recovery and resonance assignment still pose significant challenges in NMR-based complex mixture analysis. The reduced signal overlap as achieved when projecting two-dimensional (2D) J-resolved (JRES) NMR spectra can be exploited to mitigate this problem and, here, full-resolution (1)H JRES projections have been evaluated as a tool for metabolic screening and biomarker identification. We show that the recoverable information content in JRES projections is intrinsically different from that in the conventional one-dimensional (1D) and Carr-Purcell-Meiboom-Gill (CPMG) spectra, because of the combined result of reduction of the over-representation of highly split multiplet peaks and relaxation editing. Principal component and correlation analyses of full-resolution JRES spectral data demonstrated that peak alignment is necessary. The application of statistical total correlation spectroscopy (STOCSY) to JRES projections improved the identification of previously overlapped small molecule resonances in JRES (1)H NMR spectra, compared to conventional 1D and CPMG spectra. These approaches are demonstrated using a galactosamine-induced hepatotoxicity study in rats and show that JRES projections have a useful and complementary role to standard one-dimensional experiments in complex mixture analysis for improved biomarker identification.

Influenza B vaccine lineage selection--an optimized trivalent vaccine.

Epidemics of seasonal influenza viruses cause considerable morbidity and mortality each year. Various types and subtypes of influenza circulate in humans and evolve continuously such that individuals at risk of serious complications need to be vaccinated annually to keep protection up to date with circulating viruses. The influenza vaccine in most parts of the world is a trivalent vaccine, including an antigenically representative virus of recently circulating influenza A/H3N2, A/H1N1, and influenza B viruses. However, since the 1970s influenza B has split into two antigenically distinct lineages, only one of which is represented in the annual trivalent vaccine at any time. We describe a lineage selection strategy that optimizes protection against influenza B using the standard trivalent vaccine as a potentially cost effective alternative to quadrivalent vaccines.

The confounded effects of age and exposure history in response to influenza vaccination.

Numerous studies have explored whether the antibody response to influenza vaccination in elderly adults is as strong as it is in young adults. Results vary, but tend to indicate lower post-vaccination titers (antibody levels) in the elderly, supporting the concept of immunosenescence-the weakening of the immunological response related to age. Because the elderly in such studies typically have been vaccinated against influenza before enrollment, a confounding of effects occurs between age, and previous exposures, as a potential extrinsic reason for immunosenescence. We conducted a four-year study of serial annual immunizations with inactivated trivalent influenza vaccines in 136 young adults (16 to 39 years) and 122 elderly adults (62 to 92 years). Compared to data sets of previously published studies, which were designed to investigate the effect of age, this detailed longitudinal study with multiple vaccinations allowed us to also study the effect of prior vaccination history on the response to a vaccine. In response to the first vaccination, young adults produced higher post-vaccination titers, accounting for pre-vaccination titers, than elderly adults. However, upon subsequent vaccinations the difference in response to vaccination between the young and elderly age groups declined rapidly. Although age is an important factor when modeling the outcome of the first vaccination, this term lost its relevance with successive vaccinations. In fact, when we examined the data with the assumption that the elderly group had received (on average) as few as two vaccinations prior to our study, the difference due to age disappeared. Our analyses therefore show that the initial difference between the two age groups in their response to vaccination may not be uniquely explained by immunosenescence due to ageing of the immune system, but could equally be the result of the different pre-study vaccination and infection histories in the elderly.

The global antigenic diversity of swine influenza A viruses.

Swine influenza presents a substantial disease burden for pig populations worldwide and poses a potential pandemic threat to humans. There is considerable diversity in both H1 and H3 influenza viruses circulating in swine due to the frequent introductions of viruses from humans and birds coupled with geographic segregation of global swine populations. Much of this diversity is characterized genetically but the antigenic diversity of these viruses is poorly understood. Critically, the antigenic diversity shapes the risk profile of swine influenza viruses in terms of their epizootic and pandemic potential. Here, using the most comprehensive set of swine influenza virus antigenic data compiled to date, we quantify the antigenic diversity of swine influenza viruses on a multi-continental scale. The substantial antigenic diversity of recently circulating viruses in different parts of the world adds complexity to the risk profiles for the movement of swine and the potential for swine-derived infections in humans.

Dengue viruses cluster antigenically but not as discrete serotypes.

The four genetically divergent dengue virus (DENV) types are traditionally classified as serotypes. Antigenic and genetic differences among the DENV types influence disease outcome, vaccine-induced protection, epidemic magnitude, and viral evolution. We characterized antigenic diversity in the DENV types by antigenic maps constructed from neutralizing antibody titers obtained from African green monkeys and after human vaccination and natural infections. Genetically, geographically, and temporally, diverse DENV isolates clustered loosely by type, but we found that many are as similar antigenically to a virus of a different type as to some viruses of the same type. Primary infection antisera did not neutralize all viruses of the same DENV type any better than other types did up to 2 years after infection and did not show improved neutralization to homologous type isolates. That the canonical DENV types are not antigenically homogeneous has implications for vaccination and research on the dynamics of immunity, disease, and the evolution of DENV.

Quantifying the fitness advantage of polymerase substitutions in Influenza A/H7N9 viruses during adaptation to humans.

Adaptation of zoonotic influenza viruses towards efficient human-to-human transmissibility is a substantial public health concern. The recently emerged A/H7N9 influenza viruses in China provide an opportunity for quantitative studies of host-adaptation, as human-adaptive substitutions in the PB2 gene of the virus have been found in all sequenced human strains, while these substitutions have not been detected in any non-human A/H7N9 sequences. Given the currently available information, this observation suggests that the human-adaptive PB2 substitution might confer a fitness advantage to the virus in these human hosts that allows it to rise to proportions detectable by consensus sequencing over the course of a single human infection. We use a mathematical model of within-host virus evolution to estimate the fitness advantage required for a substitution to reach predominance in a single infection as a function of the duration of infection and the fraction of mutant present in the virus population that initially infects a human. The modeling results provide an estimate of the lower bound for the fitness advantage of this adaptive substitution in the currently sequenced A/H7N9 viruses. This framework can be more generally used to quantitatively estimate fitness advantages of adaptive substitutions based on the within-host prevalence of mutations. Such estimates are critical for models of cross-species transmission and host-adaptation of influenza virus infections.

The potential for respiratory droplet-transmissible A/H5N1 influenza virus to evolve in a mammalian host.

Avian A/H5N1 influenza viruses pose a pandemic threat. As few as five amino acid substitutions, or four with reassortment, might be sufficient for mammal-to-mammal transmission through respiratory droplets. From surveillance data, we found that two of these substitutions are common in A/H5N1 viruses, and thus, some viruses might require only three additional substitutions to become transmissible via respiratory droplets between mammals. We used a mathematical model of within-host virus evolution to study factors that could increase and decrease the probability of the remaining substitutions evolving after the virus has infected a mammalian host. These factors, combined with the presence of some of these substitutions in circulating strains, make a virus evolving in nature a potentially serious threat. These results highlight critical areas in which more data are needed for assessing, and potentially averting, this threat.

Non-linear modeling of 1H NMR metabonomic data using kernel-based orthogonal projections to latent structures optimized by simulated annealing.

Linear multivariate projection methods are frequently applied for predictive modeling of spectroscopic data in metabonomic studies. The OPLS method is a commonly used computational procedure for characterizing spectral metabonomic data, largely due to its favorable model interpretation properties providing separate descriptions of predictive variation and response-orthogonal structured noise. However, when the relationship between descriptor variables and the response is non-linear, conventional linear models will perform sub-optimally. In this study we have evaluated to what extent a non-linear model, kernel-based orthogonal projections to latent structures (K-OPLS), can provide enhanced predictive performance compared to the linear OPLS model. Just like its linear counterpart, K-OPLS provides separate model components for predictive variation and response-orthogonal structured noise. The improved model interpretation by this separate modeling is a property unique to K-OPLS in comparison to other kernel-based models. Simulated annealing (SA) was used for effective and automated optimization of the kernel-function parameter in K-OPLS (SA-K-OPLS). Our results reveal that the non-linear K-OPLS model provides improved prediction performance in three separate metabonomic data sets compared to the linear OPLS model. We also demonstrate how response-orthogonal K-OPLS components provide valuable biological interpretation of model and data. The metabonomic data sets were acquired using proton Nuclear Magnetic Resonance (NMR) spectroscopy, and include a study of the liver toxin galactosamine, a study of the nephrotoxin mercuric chloride and a study of Trypanosoma brucei brucei infection. Automated and user-friendly procedures for the kernel-optimization have been incorporated into version 1.1.1 of the freely available K-OPLS software package for both R and Matlab to enable easy application of K-OPLS for non-linear prediction modeling.