Automatically suggest suitable threshold for marginal regulator activities.

Description

The algorithm fits a mixture of a beta(1,beta) and beta(alpha,1) distribution to observed marginal regulator activities. Based on this mixture a cutoff is chosen such that the expected false positive rate is below a defined threshold.

Usage

suggestThreshold(prob, fpr=0.001)

Arguments

Value

a cutoff for marginal activity probabilities

Author(s)

Holger Froehlich

References

Froehlich, H. and Klau, G. (2013). Reconstructing Consensus Bayesian Network Structures with Application to Learning Molecular Interaction Networks. In: Beissbarth, T., Kollmar, M., Leha, A., Morgenstern, B., Schultz, A.-K., Waack, S., and Wingender, E., editors, Proc. German Conference on Bioinformatics, Open Access Series in Informatics, pages 46 - 55. Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik, Dagstuhl Publishing, Germany.

Examples

freq = 0.2*rbeta(100, 1, 10) + 0.8*rbeta(100, 5, 1)
thresh = suggestThreshold(freq)

Results

R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(birte)
Loading required package: RcppArmadillo
Loading required package: Rcpp
> png(filename="/home/ddbj/snapshot/RGM3/R_BC/result/birte/suggestThreshold.Rd_%03d_medium.png", width=480, height=480)
> ### Name: suggestThreshold
> ### Title: Automatically suggest suitable threshold for marginal regulator
> ###   activities.
> ### Aliases: suggestThreshold
> 
> ### ** Examples
> 
> freq = 0.2*rbeta(100, 1, 10) + 0.8*rbeta(100, 5, 1)
> thresh = suggestThreshold(freq)
start values: (alpha, beta) =  1.5 5 lambda =  0.5 0.5 
logLik =  -29.26244 (alpha, beta) =  2.497541 1.000045 lambda =  0.9269586 0.07304143 
logLik =  -29.64704 (alpha, beta) =  2.492108 1.000062 lambda =  0.9431311 0.0568689 
logLik =  -29.94668 (alpha, beta) =  2.487817 1.000062 lambda =  0.9558303 0.04416974 
logLik =  -30.17937 (alpha, beta) =  2.484462 1.000464 lambda =  0.9657573 0.03424267 
logLik =  -30.35997 (alpha, beta) =  2.48181 1.000045 lambda =  0.973492 0.02650799 
logLik =  -30.49967 (alpha, beta) =  2.479756 1.000045 lambda =  0.9795012 0.02049884 
logLik =  -30.60768 (alpha, beta) =  2.478207 1.000098 lambda =  0.9841613 0.01583868 
logLik =  -30.69114 (alpha, beta) =  2.47692 1.000055 lambda =  0.9877699 0.0122301 
logLik =  -30.75548 (alpha, beta) =  2.475996 1.001606 lambda =  0.990561 0.009439036 
logLik =  -30.80532 (alpha, beta) =  2.475178 1.000077 lambda =  0.9927188 0.007281187 
logLik =  -30.84368 (alpha, beta) =  2.474643 1.000058 lambda =  0.9943843 0.005615746 
logLik =  -30.87289 (alpha, beta) =  2.474254 1.011913 lambda =  0.9956697 0.004330297 
logLik =  -30.89586 (alpha, beta) =  2.473886 1.012285 lambda =  0.9966651 0.00333486 
logLik =  -30.91378 (alpha, beta) =  2.47359 1.000068 lambda =  0.9974322 0.002567817 
logLik =  -30.92729 (alpha, beta) =  2.473393 1.000828 lambda =  0.9980208 0.001979235 
logLik =  -30.93773 (alpha, beta) =  2.47323 1.000045 lambda =  0.9984747 0.001525344 
logLik =  -30.94566 (alpha, beta) =  2.473121 1.011945 lambda =  0.9988244 0.001175552 
logLik =  -30.95189 (alpha, beta) =  2.473029 1.01292 lambda =  0.9990951 0.0009049392 
logLik =  -30.9567 (alpha, beta) =  2.472954 1.013076 lambda =  0.9993035 0.000696531 
logLik =  -30.9604 (alpha, beta) =  2.472895 1.012503 lambda =  0.9994639 0.0005360965 
logLik =  -30.96328 (alpha, beta) =  2.472846 1.000458 lambda =  0.9995874 0.0004126298 
[1] "converged!"
> 
> 
> 
> 
> 
> dev.off()
null device 
          1 
>