Restif, O. | Search bioRxiv

bioRxiv2018

Optimizing non-invasive sampling of an infectious bat virus

Notable outbreaks of infectious viruses resulting from spillover events from bats have brought much attention to the ecological origins of bat-borne zoonoses, resulting in an increase in ecological and epidemiological studies on bat populations in Africa, Asia, and Australia. The aim of many of these studies is to identify new viral agents with field sampling methods that collect pooled urine samples from large plastic sheets placed under a bat roost. The efficiency of under-roost sampling also makes it an attractive method for gathering roost-level prevalence data. However, the method allows multiple individuals to contribute to a pooled sample, potentially introducing positive bias. To assess the ability of under-roost sampling to accurately estimate viral prevalence, we constructed a probabilistic model to explore the relationship between four sampling designs (quadrant, uniform, stratified, and random) and estimation bias. We modeled bat density and movement with a Poisson cluster process and spatial kernels, and simulated the four underroost sheet sampling designs by manipulating a spatial grid of hexagonal tiles. We performed global sensitivity analyses to identify major sources of estimation bias and provide recommendations for field studies that wish to estimate roost-level prevalence. We found that the quadrant-based design had a positive bias 5-7 times higher than other designs due to spatial auto-correlation among sampling sheets and clustering of bats in the roost. The sampling technique is therefore highly sensitive to viral presence; but lacks specificity, providing poor information regarding dynamics in viral prevalence. Given population sizes of 5000-14000, our simulation results indicate that using a stratified random design to collect 30-40 urine samples from 80-100 sheets, each with an area of 0.75-1m2, would provide sufficient estimation of true prevalence with minimum sampling bias and false negatives. However, acknowledging the general problem of data aggregation, we emphasize that robust inference of true prevalence from field data require information of underpinning roost sizes. Our findings refine our understanding of the underroost sampling technique with the aim of increasing its specificity, and suggest that the method be further developed as an efficient non-invasive sampling technique that provides roost-level estimates of viral prevalence within a bat population.

ecology↗

bioRxiv2017

Inferring Within-Host Bottleneck Size: A Bayesian Approach

A number of approaches exist for bottleneck-size estimation with respect to within-host bacterial infections; however, some are more appropriate than others under certain circumstances. A Bayesian comparison of several approaches is made in terms of the availability of isogenic multitype bacteria (e.g., WITS), knowledge of post-bottleneck dynamics, and the suitability of dilution with monotype bacteria. The results are summarised by a guiding flowchart.\n\nA sampling approach to bottleneck-size estimation is also introduced.

microbiology↗

bioRxiv2017

An Efficient Moments-Based Inference Method for Within-Host Bacterial Infection Dynamics

Over the last ten years, isogenic tagging (IT) has revolutionised the study of bacterial infection dynamics in laboratory animal models. However, quantitative analysis of IT data has been hindered by the piecemeal development of relevant statistical models. The most promising approach relies on stochastic Markovian models of bacterial population dynamics within and among organs. Here we present an efficient numerical method to fit such stochastic dynamic models to in vivo experimental IT data. A common approach to statistical inference with stochastic dynamic models relies on producing large numbers of simulations, but this remains a slow and inefficient method for all but simple problems, especially when tracking bacteria in multiple locations simultaneously. Instead, we derive and solve the systems of ordinary differential equations for the two lower-order moments of the stochastic variables (mean, variance and covariance). For any given model structure, and assuming linear dynamic rates, we demonstrate how the model parameters can be efficiently and accurately estimated by divergence minimisation. We then apply our method to an experimental dataset and compare the estimates and goodness-of-fit to those obtained by maximum likelihood estimation. While both sets of parameter estimates had overlapping confidence regions, the new method produced lower values for the division and death rates of bacteria: these improved the goodness-of-fit at the second time point at the expense of that of the first time point. This flexible framework can easily be applied to a range of experimental systems. Its computational efficiency paves the way for model comparison and optimal experimental design.\n\nAuthor SummaryRecent advancements in technology have meant that microbiologists are producing vast amounts of experimental data. However, statistical methods by which we can analyse that data, draw informative inference, and test relevant hypotheses, are much needed. Here, we present a new, efficient inference tool for estimating parameters of stochastic models, with a particular focus on models of within-host bacterial dynamics. The method relies on matching the two lower-order moments of the experimental data (i.e., mean, variance and covariance), to the moments from the mathematical model. The method is verified, and particular choices justified, through a number of simulation studies. We then use this method to estimate models that have been previously estimated using a \"gold-standard\" maximum likelihood procedure.\n\nList of symbolsO_LIA: number of animals\nC_LIO_LIT: number of tagged strains\nC_LIO_LIn: number of organs\nC_LIO_LINi: number of bacteria in organ i\nC_LIO_LImij: migration rate from organ i to organ j\nC_LIO_LIki: killing rate in organ i\nC_LIO_LIri: replication rate in organ i\nC_LIO_LI{tau}i: observation time i\nC_LIO_LIA, B, C: matrices\nC_LIO_LI{lambda}: vector of transition rates\nC_LIO_LIB: Number of bootstrap samples\nC_LIO_LI{theta}*: MDE parameter estimate\nC_LI\n\nAbbreviationsO_LIABC: approximate Bayesian computation\nC_LIO_LIIT: isogenic tagging\nC_LIO_LILV: live vaccine\nC_LIO_LIMARE: mean absolute relative error\nC_LIO_LIMDE: minimum divergence estimate\nC_LIO_LIMLE: maximum likelihood estimate\nC_LIO_LIqPCR: quantitative polymerase chain reaction\nC_LIO_LIWITS: wildtype isogenic tagged strain\nC_LI

microbiology↗