A Bayesian approach to inference about a change point model with application to DNA copy number experimental data

Jie Chen, Ayten Yiǧiter, Kuang Chao Chang

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

In this paper, we study the change-point inference problem motivated by the genomic data that were collected for the purpose of monitoring DNA copy number changes. DNA copy number changes or copy number variations (CNVs) correspond to chromosomal aberrations and signify abnormality of a cell. Cancer development or other related diseases are usually relevant to DNA copy number changes on the genome. There are inherited random noises in such data, therefore, there is a need to employ an appropriate statistical model for identifying statistically significant DNA copy number changes. This type of statistical inference is evidently crucial in cancer researches, clinical diagnostic applications, and other related genomic researches. For the high-throughput genomic data resulting from DNA copy number experiments, a mean and variance change point model (MVCM) for detecting the CNVs is appropriate.We propose to use a Bayesian approach to study the MVCM for the cases of one change and propose to use a sliding window to search for all CNVs on a given chromosome. We carry out simulation studies to evaluate the estimate of the locus of the DNA copy number change using the derived posterior probability. These simulation results show that the approach is suitable for identifying copy number changes. The approach is also illustrated on several chromosomes from nine fibroblast cancer cell line data (array-based comparative genomic hybridization data). All DNA copy number aberrations that have been identified and verified by karyotyping are detected by our approach on these cell lines.

Original languageEnglish (US)
Pages (from-to)1899-1913
Number of pages15
JournalJournal of Applied Statistics
Volume38
Issue number9
DOIs
StatePublished - Sep 1 2011
Externally publishedYes

Fingerprint

Change-point Model
Bayesian Approach
Experimental Data
Genomics
Cancer
Aberration
Chromosome
Cell
Bayesian approach
Inference
Change point
Comparative Genomics
Fibroblasts
Line
Random Noise
Sliding Window
Change Point
Posterior Probability
Statistical Inference
Statistical Model

Keywords

  • Bayesian inferences
  • CNVs
  • Change point
  • DNA copy numbers
  • Non-informative priors

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

A Bayesian approach to inference about a change point model with application to DNA copy number experimental data. / Chen, Jie; Yiǧiter, Ayten; Chang, Kuang Chao.

In: Journal of Applied Statistics, Vol. 38, No. 9, 01.09.2011, p. 1899-1913.

Research output: Contribution to journalArticle

@article{e4c2678e748048ccb52a479d3310795a,
title = "A Bayesian approach to inference about a change point model with application to DNA copy number experimental data",
abstract = "In this paper, we study the change-point inference problem motivated by the genomic data that were collected for the purpose of monitoring DNA copy number changes. DNA copy number changes or copy number variations (CNVs) correspond to chromosomal aberrations and signify abnormality of a cell. Cancer development or other related diseases are usually relevant to DNA copy number changes on the genome. There are inherited random noises in such data, therefore, there is a need to employ an appropriate statistical model for identifying statistically significant DNA copy number changes. This type of statistical inference is evidently crucial in cancer researches, clinical diagnostic applications, and other related genomic researches. For the high-throughput genomic data resulting from DNA copy number experiments, a mean and variance change point model (MVCM) for detecting the CNVs is appropriate.We propose to use a Bayesian approach to study the MVCM for the cases of one change and propose to use a sliding window to search for all CNVs on a given chromosome. We carry out simulation studies to evaluate the estimate of the locus of the DNA copy number change using the derived posterior probability. These simulation results show that the approach is suitable for identifying copy number changes. The approach is also illustrated on several chromosomes from nine fibroblast cancer cell line data (array-based comparative genomic hybridization data). All DNA copy number aberrations that have been identified and verified by karyotyping are detected by our approach on these cell lines.",
keywords = "Bayesian inferences, CNVs, Change point, DNA copy numbers, Non-informative priors",
author = "Jie Chen and Ayten Yiǧiter and Chang, {Kuang Chao}",
year = "2011",
month = "9",
day = "1",
doi = "10.1080/02664763.2010.529886",
language = "English (US)",
volume = "38",
pages = "1899--1913",
journal = "Journal of Applied Statistics",
issn = "0266-4763",
publisher = "Routledge",
number = "9",

}

TY - JOUR

T1 - A Bayesian approach to inference about a change point model with application to DNA copy number experimental data

AU - Chen, Jie

AU - Yiǧiter, Ayten

AU - Chang, Kuang Chao

PY - 2011/9/1

Y1 - 2011/9/1

N2 - In this paper, we study the change-point inference problem motivated by the genomic data that were collected for the purpose of monitoring DNA copy number changes. DNA copy number changes or copy number variations (CNVs) correspond to chromosomal aberrations and signify abnormality of a cell. Cancer development or other related diseases are usually relevant to DNA copy number changes on the genome. There are inherited random noises in such data, therefore, there is a need to employ an appropriate statistical model for identifying statistically significant DNA copy number changes. This type of statistical inference is evidently crucial in cancer researches, clinical diagnostic applications, and other related genomic researches. For the high-throughput genomic data resulting from DNA copy number experiments, a mean and variance change point model (MVCM) for detecting the CNVs is appropriate.We propose to use a Bayesian approach to study the MVCM for the cases of one change and propose to use a sliding window to search for all CNVs on a given chromosome. We carry out simulation studies to evaluate the estimate of the locus of the DNA copy number change using the derived posterior probability. These simulation results show that the approach is suitable for identifying copy number changes. The approach is also illustrated on several chromosomes from nine fibroblast cancer cell line data (array-based comparative genomic hybridization data). All DNA copy number aberrations that have been identified and verified by karyotyping are detected by our approach on these cell lines.

AB - In this paper, we study the change-point inference problem motivated by the genomic data that were collected for the purpose of monitoring DNA copy number changes. DNA copy number changes or copy number variations (CNVs) correspond to chromosomal aberrations and signify abnormality of a cell. Cancer development or other related diseases are usually relevant to DNA copy number changes on the genome. There are inherited random noises in such data, therefore, there is a need to employ an appropriate statistical model for identifying statistically significant DNA copy number changes. This type of statistical inference is evidently crucial in cancer researches, clinical diagnostic applications, and other related genomic researches. For the high-throughput genomic data resulting from DNA copy number experiments, a mean and variance change point model (MVCM) for detecting the CNVs is appropriate.We propose to use a Bayesian approach to study the MVCM for the cases of one change and propose to use a sliding window to search for all CNVs on a given chromosome. We carry out simulation studies to evaluate the estimate of the locus of the DNA copy number change using the derived posterior probability. These simulation results show that the approach is suitable for identifying copy number changes. The approach is also illustrated on several chromosomes from nine fibroblast cancer cell line data (array-based comparative genomic hybridization data). All DNA copy number aberrations that have been identified and verified by karyotyping are detected by our approach on these cell lines.

KW - Bayesian inferences

KW - CNVs

KW - Change point

KW - DNA copy numbers

KW - Non-informative priors

UR - http://www.scopus.com/inward/record.url?scp=79961162685&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79961162685&partnerID=8YFLogxK

U2 - 10.1080/02664763.2010.529886

DO - 10.1080/02664763.2010.529886

M3 - Article

VL - 38

SP - 1899

EP - 1913

JO - Journal of Applied Statistics

JF - Journal of Applied Statistics

SN - 0266-4763

IS - 9

ER -