Variable selection in high-dimensional multivariate binary data with application to the analysis of microbial community DNA fingerprints

Soil features in rookeries of Antarctic penguins reveal sea to land biotransport of chemical pollutants

The main soil physical-chemical features, the concentrations of a set of pollutants, and the soil microbiota linked to penguin rookeries have been studied in 10 selected sites located at the South Shetland Islands and the Antarctic Peninsula (Maritime Antarctica). This study aims to test the hypothesis that biotransport by penguins increases the concentration of pollutants, especially heavy metals, in Antarctic soils, and alters its microbiota. Our results show that penguins do transport certain chemical elements and thus cause accumulation in land areas through their excreta. Overall, a higher penguin activity is associated with higher organic carbon content and with higher concentrations of certain pollutants in soils, especially cadmium, cooper and arsenic, as well as zinc and selenium. In contrast, in soils that are less affected by penguins’ faecal depositions, the concentrations of elements of geochemical origin, such as iron and cobalt, increase their relative weighted contribution, whereas the above-mentioned pollutants maintain very low levels. The concentrations of pollutants are far higher in those penguin rookeries that are more exposed to ship traffic. In addition, the soil microbiota of penguin-influenced soils was studied by molecular methods. Heavily penguin-affected soils have a massive presence of enteric bacteria, whose relative dominance can be taken as an indicator of penguin influence. Faecal bacteria are present in addition to typical soil taxa, the former becoming dominant in the microbiota of penguin-affected soils, whereas typical soil bacteria, such as Actinomycetales, co-dominate the microbiota of less affected soils. Results indicate that the continuous supply by penguin faeces, and not the selectivity by increased pollutant concentrations is the main factor shaping the soil bacterial community. Overall, massive penguin influence results in increased concentrations of certain pollutants and in a strong change in taxa dominance in the soil bacterial community.

The effect of heavy metal contamination on the bacterial community structure at Jiaozhou Bay, China

In this study, determination of heavy metal parameters and microbiological characterization of marine sediments obtained from two heavily polluted sites and one low-grade contaminated reference station at Jiaozhou Bay in China were carried out. The microbial communities found in the sampled marine sediments were studied using PCR-DGGE (denaturing gradient gel electrophoresis) fingerprinting profiles in combination with multivariate analysis. Clustering analysis of DGGE and matrix of heavy metals displayed similar occurrence patterns. On this basis, 17 samples were classified into two clusters depending on the presence or absence of the high level contamination. Moreover, the cluster of highly contaminated samples was further classified into two sub-groups based on the stations of their origin. These results showed that the composition of the bacterial community is strongly influenced by heavy metal variables present in the sediments found in the Jiaozhou Bay. This study also suggested that metagenomic techniques such as PCR-DGGE fingerprinting in combination with multivariate analysis is an efficient method to examine the effect of metal contamination on the bacterial community structure.

Effects of dietary resistant starch content on metabolic status, milk composition, and microbial profiling in lactating sows and on offspring performance

In the present study, the effects of dietary resistant starch (RS) content on serum metabolite and hormone concentrations, milk composition, and faecal microbial profiling in lactating sows, as well as on offspring performance was investigated. Sixteen sows were randomly allotted at breeding to two treatments containing low- and high-RS contents from normal and high-amylose corn varieties, respectively, and each treatment had eight replicates (sows). Individual piglet body weight (BW) and litter size were recorded at birth and weaning. Milk samples were obtained on day 10 after farrowing for composition analysis. On day 2 before weaning, blood and faecal samples were collected to determine serum metabolite and hormone concentrations and faecal microbial populations, respectively. Litter size at birth and weaning were not influenced (p > 0.05) by the sow dietary treatments. Although feeding the RS-rich diet to sows reduced (p = 0.004) offspring birth BW, there was no difference in piglet BW at weaning (p > 0.05). High-RS diet increased (p < 0.05) serum triacylglycerol and nonesterified fatty acid concentrations and milk total solid content, and tended (p = 0.09) to increase milk fat content in lactating sows. Feeding the RS-rich diet to sows increased (p < 0.01) faecal bacterial population diversity. These results indicate that high-RS diets induce fatty acid mobilization and a greater intestinal bacterial richness in lactating sows, as well as a greater nutrient density in maternal milk, without affecting offspring performance at weaning.

Differential protein expression and peak selection in mass spectrometry data by binary discriminant analysis

MOTIVATION

Proteomic mass spectrometry analysis is becoming routine in clinical diagnostics, for example to monitor cancer biomarkers using blood samples. However, differential proteomics and identification of peaks relevant for class separation remains challenging.

RESULTS

Here, we introduce a simple yet effective approach for identifying differentially expressed proteins using binary discriminant analysis. This approach works by data-adaptive thresholding of protein expression values and subsequent ranking of the dichotomized features using a relative entropy measure. Our framework may be viewed as a generalization of the 'peak probability contrast' approach of Tibshirani et al. (2004) and can be applied both in the two-group and the multi-group setting. Our approach is computationally inexpensive and shows in the analysis of a large-scale drug discovery test dataset equivalent prediction accuracy as a random forest. Furthermore, we were able to identify in the analysis of mass spectrometry data from a pancreas cancer study biological relevant and statistically predictive marker peaks unrecognized in the original study.

AVAILABILITY AND IMPLEMENTATION

The methodology for binary discriminant analysis is implemented in the R package binda, which is freely available under the GNU General Public License (version 3 or later) from CRAN at URL http://cran.r-project.org/web/packages/binda/. R scripts reproducing all described analyzes are available from the web page http://strimmerlab.org/software/binda/.

CONTACT

k.strimmer@imperial.ac.uk.

Dietary fat content and fiber type modulate hind gut microbial community and metabolic markers in the pig

Obesity leads to changes in the gut microbial community which contribute to the metabolic dysregulation in obesity. Dietary fat and fiber affect the caloric density of foods. The impact of dietary fat content and fiber type on the microbial community in the hind gut is unknown. Effect of dietary fat level and fiber type on hindgut microbiota and volatile fatty acid (VFA) profiles was investigated. Expression of metabolic marker genes in the gut, adipose tissue and liver was determined. A 2×2 experiment was conducted in pigs fed at two dietary fat levels (5% or 17.5% swine grease) and two fiber types (4% inulin, fermentable fructo-oligosaccharide or 4% solka floc, non-fermentable cellulose). High fat diets (HFD) resulted in a higher (P<0.05) total body weight gain, feed efficiency and back fat accumulation than the low fat diet. Feeding of inulin, but not solka floc, attenuated (P<0.05) the HFD-induced higher body weight gain and fat mass accumulation. Inulin feeding tended to lead to higher total VFA production in the cecum and resulted in a higher (P<0.05) expression of acyl coA oxidase (ACO), a marker of peroxisomal β-oxidation. Inulin feeding also resulted in lower expression of sterol regulatory element binding protein 1c (SREBP-1c), a marker of lipid anabolism. Bacteria community structure characterized by DGGE analysis of PCR amplified 16S rRNA gene fragments showed that inulin feeding resulted in greater bacterial population richness than solka floc feeding. Cluster analysis of pairwise Dice similarity comparisons of the DGGE profiles showed grouping by fiber type but not the level of dietary fat. Canonical correspondence analysis (CCA) of PCR- DGGE profiles showed that inulin feeding negatively correlated with back fat thickness. This study suggests a strong interplay between dietary fat level and fiber type in determining susceptibility to obesity.