Model fits to the observed data. Posterior mean (red), 50% (dark blue) and 95% credible (light blue) intervals for θt in terms of log transformed proportions (black). The raw count data are shown in black for comparison. A pseudo-count of 0.65 was added to the raw data prior to normalization and log-transformation to avoid taking the log of zero values. (PDF 163 kb
Summary of the Kalman Filter, Kalman Smoother, and Backwards Sampling algorithm. (PDF 295 kb
Posterior 95% credible regions for bacterial dynamics (θ) in the PhILR Basis. (Left) The hierarchica...
Artificial gut setup. (1) Reactor vessels; (2) flow meters controlling gas inputs; (3) pump-fed medi...
Dynamics inferred in the balance between the Enterobacteriaceae and all other taxa does not correlat...
Proportions of most abundant bacterial families estimated from count data during hourly sampling per...
An inverse relationship between biological variation and initial relative abundance. 5, 25, 50, 75, ...
Table of parameters in MALLARD model used to analyze artificial gut dataset. For the artificial gut ...
Analysis of a toy simulated microbial community demonstrates the advantages of accounting for techni...
Posterior estimates for the regression slope between biological variation and bacterial family start...
Posterior estimates for the percent of total variation attributable to biological sources is not sen...
Four bacterial families comprise most the biological variation in our study. Relative biological var...
Irregular sub-daily oscillation observed in PhILR balance n12 does not correlate with known external...
Samples were collected over a one month period with both daily and hourly sampling intervals. Daily ...
The decomposition of technical variation among bacterial families. Posterior distribution for log-ra...
PCoA based on Aitchison distance applied to most abundant bacterial families. To avoid taking the lo...
Summary of the Kalman Filter, Kalman Smoother, and Backwards Sampling algorithm. (PDF 295 kb
Posterior 95% credible regions for bacterial dynamics (θ) in the PhILR Basis. (Left) The hierarchica...
Artificial gut setup. (1) Reactor vessels; (2) flow meters controlling gas inputs; (3) pump-fed medi...
Dynamics inferred in the balance between the Enterobacteriaceae and all other taxa does not correlat...
Proportions of most abundant bacterial families estimated from count data during hourly sampling per...
An inverse relationship between biological variation and initial relative abundance. 5, 25, 50, 75, ...
Table of parameters in MALLARD model used to analyze artificial gut dataset. For the artificial gut ...
Analysis of a toy simulated microbial community demonstrates the advantages of accounting for techni...
Posterior estimates for the regression slope between biological variation and bacterial family start...
Posterior estimates for the percent of total variation attributable to biological sources is not sen...
Four bacterial families comprise most the biological variation in our study. Relative biological var...
Irregular sub-daily oscillation observed in PhILR balance n12 does not correlate with known external...
Samples were collected over a one month period with both daily and hourly sampling intervals. Daily ...
The decomposition of technical variation among bacterial families. Posterior distribution for log-ra...
PCoA based on Aitchison distance applied to most abundant bacterial families. To avoid taking the lo...
Summary of the Kalman Filter, Kalman Smoother, and Backwards Sampling algorithm. (PDF 295 kb
Posterior 95% credible regions for bacterial dynamics (θ) in the PhILR Basis. (Left) The hierarchica...
Artificial gut setup. (1) Reactor vessels; (2) flow meters controlling gas inputs; (3) pump-fed medi...