1887
banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
Efficient Bayesian estimation of Markov model transition matrices with given stationary distribution
Rent:
Rent this article for
USD
10.1063/1.4801325
/content/aip/journal/jcp/138/16/10.1063/1.4801325
http://aip.metastore.ingenta.com/content/aip/journal/jcp/138/16/10.1063/1.4801325

Figures

Image of FIG. 1.
FIG. 1.

The conditional density p(x|a, b, c, d) for a = 8.0, b = 2.0, c = 4.0, and d = 30.0 (solid line) and the corresponding enveloping function g(x) (dashed line). The density was scaled to the mode point value p(x m |a, b, c, d) to fit it into the range [0, 1].

Image of FIG. 2.
FIG. 2.

The picture shows the location of the optimal points and for p(x) (solid line) and the points x l and x u obtained from the Gaussian approximation (dashed line). The Gaussian approximation touches the density at the mode point.

Image of FIG. 3.
FIG. 3.

Conditional density p(x) (solid line) for different parameters a, b, c, and d. The histograms show a sample of N = 106 random variates generated using the method outlined above. First row c = 4, d = 10: (a) a = 5, b = 0, (b) a = 5, b = 2, and (c) a = 0, b = 2. Second row c = 40, d = 100: (d) a = 100, b = 5, (e) a = 100, b = 100, and (f) a = 5, b = 100. Third row: (g) a = 0, b = 0, c = 4, d = 10, (h) a = 0.5, b = 0.2, c = 40, d = 100, and (i) a = 0, b = 0, c = 4 × 103, d = 104.

Image of FIG. 4.
FIG. 4.

Results obtained for the model system with count matrix (23) and stationary distribution (24) . Standard deviation for estimated mean and variance of observables is plotted against the number of elementary sampling steps N. (a) Mean transition matrix element , (b) mean of the second largest implied time scale , (c) transition matrix element variance , and (d) variance of the second largest implied time scale . The Gibbs sampler introduced here (solid line) convergences faster than the Metropolis chain from Ref. 54 (dashed line) by almost two orders of magnitude. For the mean second largest time scale , (b), the achieved speedup is more than one order of magnitude.

Image of FIG. 5.
FIG. 5.

Autocorrelation functions for the model system with count matrix (23) and stationary distribution (24) . (a) Autocorrelation function for the transition matrix element p 13 and (b) autocorrelation function for the second largest implied time scale t 2. The number of steps to take until samples are decorrelated n decorr is two orders of magnitude smaller for the Gibbs sampling method (solid line), n decorr = 3 for p 13 as well as for t 2, compared to the Metropolis sampling method (dashed line), n decorr = 123 for p 13 and n decorr = 135 for t 2.

Image of FIG. 6.
FIG. 6.

Results obtained for the synthetic peptide MR121-GSGS-W. Standard deviation of Figure 6(a) the mean implied time scale and Figure 6(b) the implied time scale variance . The Gibbs sampler (solid line) shows a faster convergence than the Metropolis sampler (dashed line) for mean and variance of the second largest implied timescale t 2.

Image of FIG. 7.
FIG. 7.

Autocorrelation function for the MR121-GSGS-W peptide count matrix. The number of steps to take until samples are decorrelated n decorr is an order of magnitude smaller for the Gibbs sampling method (solid line), n decorr = 4600, compared to the Metropolis sampling method (dashed line), n decorr = 33 000.

Image of FIG. 8.
FIG. 8.

Histograms for implied time scale t 2 corresponding to second largest eigenvalue λ2. The histogram of timescales generated by incorporating knowledge about stationary probabilities (a) in comparison to the histogram of timescales generated by the non-reversible (b) and reversible method (c). It is clearly visible that the sample mean (dashed line) gives a more accurate prediction of the true value (solid line) due to the additional information about stationary probabilities.

Tables

Generic image for table

Sample p(x|a, b, c, d).

Generic image for table

Modified rejection algorithm.

Generic image for table

Iterative maximum likelihood estimation with fixed stationary distribution.

Generic image for table

Gibbs sampling of P with fixed stationary distribution π.

Generic image for table

Gibbs sampling of P with uncertain stationary distribution.

Generic image for table

Gibbs sampling of P with fixed stationary distribution π—Sparse version.

Generic image for table
Table I.

Decorrelation times for estimated mean and variances of observables. The results were generated for the model system with count matrix (23) and stationary distribution (24) (p 13 and t 2) and for the MR121-GSGS-W peptide (t 2 only).

Loading

Article metrics loading...

/content/aip/journal/jcp/138/16/10.1063/1.4801325
2013-04-24
2014-04-18
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: Efficient Bayesian estimation of Markov model transition matrices with given stationary distribution
http://aip.metastore.ingenta.com/content/aip/journal/jcp/138/16/10.1063/1.4801325
10.1063/1.4801325
SEARCH_EXPAND_ITEM