Supplementary Data; Mitchell et al.Reduced Intestinal Motility, Mucosal Barrier Function, and Inflammation in Aged Monkeys
Methods
Fecal DNA libraries were sequenced on an IlluminaHiSeq2000 . Raw reads were then processed as previously described (1, 2) and were fed into AbundantOTU+ v.0.93b ( with the “-abundantonly” option to pick de novo Operational Taxonomic Units (OTU). AbundantOTU+ generated 1,327 OTUs, incorporating 35,742,822 (99.60%) of all the input reads and discarded those that were not incorporated. UCHIME ( in combination with the Gold reference database were used to screen for chimeras in OUT sequences and none were found.
Taxonomic assignments were done by aligning the OTU sequences to Silva database (release 108, using BLASTn v. 2.2.26+ (expectation value set to e-5). After that, RDP classifier v. 2.5 was utilized to classify the full-length Silva sequences that returned the best BLASTn match to our OTUs. We only considered taxonomic assignments from RDP with confidence score ≥ 80% (3). Results were organized into tables where samples were rows and each OTU raw counts were in columns. The raw counts were then normalized and log transformed as follows:
The normalized and log10 transformed counts were used to generate Bray-Curtis dissimilarity matrix, and to perform a multidimensional scaling (MDS) using the function “capscale” in the R package “vegan”.
In our mixed linear model, we utilized a random term with each individual animal nested within cage and a fixed term for age of the vervet and the number of days from the beginning of our study. Because our animals were on distinct experimental diets (4), we built separate statistical models for monkeys that were on chow diet versus vs. those on a diet high in fructose with the following code in R.
We compared the two models using the “anova” function in R to obtain p-values for the animal nested within cage and age coefficients within our model. We also built a model on all of the data adding a fixed term for diet. Eight of our 52 samples were collected weeks before the experiment. Consistent with drift in the microbial community over short periods of time within the vervet colony, these samples were significant outliers when compared to our other samples collected between days 34 and 69 of our study. These baseline samples were therefore removed from our statistical models.
Scripts that we used are available here:
Supplementary Table 1
The detection frequency ofintestinal mucosal permeability marker, FITC-Dextran 40, in plasma samples. Three young and four old monkeys were sampled after oral dosing.
Day 1 / Day 2 / Day 3 / Overall frequency p-valueYoung / 0/3 / 0/3 / 0/3
Old / 2/4 / 2/4 / 1/4 / 0.039
Supplementary Table 2
Regression results for microbiome analysis of monkeys eating a chow diet with age included as fixed term and animal nested within cage as a random term. P-values shown are for significance of the coefficients of the linear model for the first two MDS axes at the family level.
MDS Axis / Percent variation explained / Agep-value / Days
p-value / Cage/Animal effect
p-value
1 / 43.19% / 0.830 / 0.027 / 1.01 x 10-10
2 / 15.91% / 0.733 / 0.002 / 1.41 x 10-8
Supplementary Table 3
Regression results for microbiome analysis of monkeys eating a high fructose diet with age included as fixed term and animal nested within cage as a random term. P-values shown are for significance of the coefficients of the linear model for the first two MDS axes at the family level. from the standard linear model are shown and a model without the assumption of sample independence, are shown. We also show p-values for effects of time, the individual monkey and housing.
MDS Axis / Percent variation explained / Agep-value / Days
p-value / Cage/Animal effect
p-value
1 / 28.21% / 0.289 / 3.52 x 10-5 / 6.58 x 10-8
2 / 20.93% / 0.807 / 0.031 / 9.34 x 10-5
References Cited
1.Arthur JC, Perez-Chanona E, Muhlbauer M, Tomkovich S, Uronis JM, Fan TJ, et al. Intestinal inflammation targets cancer-inducing activity of the microbiota.Science.2012;338:120-123.
2.McCafferty J, Muhlbauer M, Gharaibeh RZ, Arthur JC, Perez-Chanona E, Sha W, et al. Stochastic changes over time and not founder effects drive cage effects in microbial community assembly in a mouse model.The ISME journal.2013;7:2116-2125.
3.Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy.Applied and environmental microbiology.2007;73:5261-5267.
4.Kavanagh K, Wylie AT, Tucker KL, Hamp TJ, Gharaibeh RZ, Fodor AA, et al. Dietary fructose induces endotoxemia and hepatic injury in calorically controlled primates.American Journal of Clinical Nutrition.2013;98:349-357.