The key operation in Bayesian inference is to compute high-dimensional integrals. An old approximate technique is the Laplace method or approximation, which dates back to Pierre-Simon Laplace (1774). This simple idea approximates the integrand with a second-order Taylor expansion around the mode and computes the integral analytically. By developing a nested version of this classical idea, combined with modern numerical techniques for sparse matrices, we obtain the approach of integrated nested Laplace approximations (INLA) to do approximate Bayesian inference for latent Gaussian models (LGMs). LGMs represent an important model abstraction for Bayesian inference and include a large proportion of the statistical models used today. In this review, we discuss the reasons for the success of the INLA approach, the R-INLA package, why it is so accurate, why the approximations are very quick to compute, and why LGMs make such a useful concept for Bayesian computing.

/pdf/bayesian-computing-with-inla-a-review-179hbvrdx8.pdf

Bayesian Computing with INLA: A Review

The human gut microbiota is impacted by host nutrition and health status and therefore represents a potentially adaptive phenotype influenced by metabolic and immune constraints. Previous studies contrasting rural populations in developing countries to urban industrialized ones have shown that industrialization is strongly correlated with patterns in human gut microbiota; however, we know little about the relative contribution of factors such as climate, diet, medicine, hygiene practices, host genetics, and parasitism. Here, we focus on fine-scale comparisons of African rural populations in order to (i) contrast the gut microbiota of populations inhabiting similar environments but having different traditional subsistence modes and either shared or distinct genetic ancestry, and (ii) examine the relationship between gut parasites and bacterial communities. Characterizing the fecal microbiota of Pygmy hunter-gatherers as well as Bantu individuals from both farming and fishing populations in Southwest Cameroon, we found that the gut parasite Entamoeba is significantly correlated with microbiome composition and diversity. We show that across populations, colonization by this protozoa can be predicted with 79% accuracy based on the composition of an individual's gut microbiota, and that several of the taxa most important for distinguishing Entamoeba absence or presence are signature taxa for autoimmune disorders. We also found gut communities to vary significantly with subsistence mode, notably with some taxa previously shown to be enriched in other hunter-gatherers groups (in Tanzania and Peru) also discriminating hunter-gatherers from neighboring farming or fishing populations in Cameroon.

/pdf/variation-in-rural-african-gut-microbiota-is-strongly-3abp5m1h7e.pdf

Variation in Rural African Gut Microbiota Is Strongly Correlated with Colonization by Entamoeba and Subsistence

In 2001, the World Health Assembly (WHA) passed the landmark WHA 54.19 resolution for global scale-up of mass administration of anthelmintic drugs for morbidity control of schistosomiasis and soil-transmitted helminthiasis, which affect more than 1·5 billion of the world's poorest people. Since then, more than a decade of research and experience has yielded crucial knowledge on the control and elimination of these helminthiases. However, the global strategy has remained largely unchanged since the original 2001 WHA resolution and associated WHO guidelines on preventive chemotherapy. In this Personal View, we highlight recent advances that, taken together, support a call to revise the global strategy and guidelines for preventive chemotherapy and complementary interventions against schistosomiasis and soil-transmitted helminthiasis. These advances include the development of guidance that is specific to goals of morbidity control and elimination of transmission. We quantify the result of forgoing this opportunity by computing the yearly disease burden, mortality, and lost economic productivity associated with maintaining the status quo. Without change, we estimate that the population of sub-Saharan Africa will probably lose 2·3 million disability-adjusted life-years and US$3·5 billion of economic productivity every year, which is comparable to recent acute epidemics, including the 2014 Ebola and 2015 Zika epidemics. We propose that the time is now to strengthen the global strategy to address the substantial disease burden of schistosomiasis and soil-transmitted helminthiasis.

A Call to Strengthen the Global Strategy Against Schistosomiasis and Soil-Transmitted Helminthiasis: The Time Is Now

The key operation in Bayesian inference, is to compute high-dimensional integrals. An old approximate technique is the Laplace method or approximation, which dates back to Pierre- Simon Laplace (1774). This simple idea approximates the integrand with a second order Taylor expansion around the mode and computes the integral analytically. By developing a nested version of this classical idea, combined with modern numerical techniques for sparse matrices, we obtain the approach of Integrated Nested Laplace Approximations (INLA) to do approximate Bayesian inference for latent Gaussian models (LGMs). LGMs represent an important model-abstraction for Bayesian inference and include a large proportion of the statistical models used today. In this review, we will discuss the reasons for the success of the INLA-approach, the R-INLA package, why it is so accurate, why the approximations are very quick to compute and why LGMs make such a useful concept for Bayesian computing.

https://thelancet.com/pdfs/journals/langlo/PIIS2214-109X(15)00047-9.pdf

Comparison of community-wide, integrated mass drug administration strategies for schistosomiasis and soil-transmitted helminthiasis: a cost-effectiveness modelling study

Summary Background Interest is growing in predictive risk mapping for neglected tropical diseases (NTDs), particularly to scale up preventive chemotherapy, surveillance, and elimination efforts. Soil-transmitted helminths (hookworm, Ascaris lumbricoides , and Trichuris trichiura ) are the most widespread NTDs, but broad geographical analyses are scarce. We aimed to predict the spatial and temporal distribution of soil-transmitted helminth infections, including the number of infected people and treatment needs, across sub-Saharan Africa. Methods We systematically searched PubMed, Web of Knowledge, and African Journal Online from inception to Dec 31, 2013, without language restrictions, to identify georeferenced surveys. We extracted data from household surveys on sources of drinking water, sanitation, and women's level of education. Bayesian geostatistical models were used to align the data in space and estimate risk of with hookworm, A lumbricoides , and T trichiura over a grid of roughly 1 million pixels at a spatial resolution of 5 × 5 km. We calculated anthelmintic treatment needs on the basis of WHO guidelines (treatment of all school-aged children once per year where prevalence in this population is 20–50% or twice per year if prevalence is greater than 50%). Findings We identified 459 relevant survey reports that referenced 6040 unique locations. We estimate that the prevalence of hookworm, A lumbricoides , and T trichiura among school-aged children from 2000 onwards was 16·5%, 6·6%, and 4·4%. These estimates are between 52% and 74% lower than those in surveys done before 2000, and have become similar to values for the entire communities. We estimated that 126 million doses of anthelmintic treatments are required per year. Interpretation Patterns of soil-transmitted helminth infection in sub-Saharan Africa have changed and the prevalence of infection has declined substantially in this millennium, probably due to socioeconomic development and large-scale deworming programmes. The global control strategy should be reassessed, with emphasis given also to adults to progress towards local elimination. Funding Swiss National Science Foundation and European Research Council.

https://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099(14)71004-7.pdf

Erika Langer

Papers

Spatial and temporal distribution of soil-transmitted helminth infection in sub-Saharan Africa: a systematic review and geostatistical meta-analysis