Significance evaluation in factor graphs

Publication: Research - peer-reviewJournal article

Standard

Significance evaluation in factor graphs. / Madsen, Tobias; Hobolth, Asger; Jensen, Jens Ledet; Pedersen, Jakob Skou.

In: BMC Bioinformatics , Vol. 18, 31.03.2017, p. 199.

Publication: Research - peer-reviewJournal article

Harvard

APA

CBE

MLA

Madsen, Tobias et al."Significance evaluation in factor graphs". BMC Bioinformatics . 2017, 18. 199. Available: 10.1186/s12859-017-1614-z

Vancouver

Author

Madsen, Tobias; Hobolth, Asger; Jensen, Jens Ledet; Pedersen, Jakob Skou / Significance evaluation in factor graphs.

In: BMC Bioinformatics , Vol. 18, 31.03.2017, p. 199.

Publication: Research - peer-reviewJournal article

Bibtex

@article{76095c473b604a4aa8f98ca697b060f1,
title = "Significance evaluation in factor graphs",
abstract = "BackgroundFactor graphs provide a flexible and general framework for specifying probability distributions. They can capture a range of popular and recent models for analysis of both genomics data as well as data from other scientific fields. Owing to the ever larger data sets encountered in genomics and the multiple-testing issues accompanying them, accurate significance evaluation is of great importance. We here address the problem of evaluating statistical significance of observations from factor graph models.ResultsTwo novel numerical approximations for evaluation of statistical significance are presented. First a method using importance sampling. Second a saddlepoint approximation based method. We develop algorithms to efficiently compute the approximations and compare them to naive sampling and the normal approximation. The individual merits of the methods are analysed both from a theoretical viewpoint and with simulations. A guideline for choosing between the normal approximation, saddle-point approximation and importance sampling is also provided. Finally, the applicability of the methods is demonstrated with examples from cancer genomics, motif-analysis and phylogenetics.ConclusionsThe applicability of saddlepoint approximation and importance sampling is demonstrated on known models in the factor graph framework. Using the two methods we can substantially improve computational cost without compromising accuracy. This contribution allows analyses of large datasets in the general factor graph framework.",
keywords = "Significance evaluation, factor graph, saddlepoint approximation, importance sampling",
author = "Tobias Madsen and Asger Hobolth and Jensen, {Jens Ledet} and Pedersen, {Jakob Skou}",
year = "2017",
month = "3",
doi = "10.1186/s12859-017-1614-z",
volume = "18",
pages = "199",
journal = "B M C Bioinformatics",
issn = "1471-2105",
publisher = "BioMed Central Ltd.",

}

RIS

TY - JOUR

T1 - Significance evaluation in factor graphs

AU - Madsen,Tobias

AU - Hobolth,Asger

AU - Jensen,Jens Ledet

AU - Pedersen,Jakob Skou

PY - 2017/3/31

Y1 - 2017/3/31

N2 - BackgroundFactor graphs provide a flexible and general framework for specifying probability distributions. They can capture a range of popular and recent models for analysis of both genomics data as well as data from other scientific fields. Owing to the ever larger data sets encountered in genomics and the multiple-testing issues accompanying them, accurate significance evaluation is of great importance. We here address the problem of evaluating statistical significance of observations from factor graph models.ResultsTwo novel numerical approximations for evaluation of statistical significance are presented. First a method using importance sampling. Second a saddlepoint approximation based method. We develop algorithms to efficiently compute the approximations and compare them to naive sampling and the normal approximation. The individual merits of the methods are analysed both from a theoretical viewpoint and with simulations. A guideline for choosing between the normal approximation, saddle-point approximation and importance sampling is also provided. Finally, the applicability of the methods is demonstrated with examples from cancer genomics, motif-analysis and phylogenetics.ConclusionsThe applicability of saddlepoint approximation and importance sampling is demonstrated on known models in the factor graph framework. Using the two methods we can substantially improve computational cost without compromising accuracy. This contribution allows analyses of large datasets in the general factor graph framework.

AB - BackgroundFactor graphs provide a flexible and general framework for specifying probability distributions. They can capture a range of popular and recent models for analysis of both genomics data as well as data from other scientific fields. Owing to the ever larger data sets encountered in genomics and the multiple-testing issues accompanying them, accurate significance evaluation is of great importance. We here address the problem of evaluating statistical significance of observations from factor graph models.ResultsTwo novel numerical approximations for evaluation of statistical significance are presented. First a method using importance sampling. Second a saddlepoint approximation based method. We develop algorithms to efficiently compute the approximations and compare them to naive sampling and the normal approximation. The individual merits of the methods are analysed both from a theoretical viewpoint and with simulations. A guideline for choosing between the normal approximation, saddle-point approximation and importance sampling is also provided. Finally, the applicability of the methods is demonstrated with examples from cancer genomics, motif-analysis and phylogenetics.ConclusionsThe applicability of saddlepoint approximation and importance sampling is demonstrated on known models in the factor graph framework. Using the two methods we can substantially improve computational cost without compromising accuracy. This contribution allows analyses of large datasets in the general factor graph framework.

KW - Significance evaluation

KW - factor graph

KW - saddlepoint approximation

KW - importance sampling

U2 - 10.1186/s12859-017-1614-z

DO - 10.1186/s12859-017-1614-z

M3 - Journal article

VL - 18

SP - 199

JO - B M C Bioinformatics

T2 - B M C Bioinformatics

JF - B M C Bioinformatics

SN - 1471-2105

ER -