A wavelet-based method to exploit epigenomic language in the regulatory region
Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Standard
A wavelet-based method to exploit epigenomic language in the regulatory region. / Nguyen, Nha; Vo, An; Won, Kyoung-Jae.
I: Bioinformatics (Online), Bind 30, Nr. 7, 01.04.2014, s. 908-14.Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - A wavelet-based method to exploit epigenomic language in the regulatory region
AU - Nguyen, Nha
AU - Vo, An
AU - Won, Kyoung-Jae
PY - 2014/4/1
Y1 - 2014/4/1
N2 - MOTIVATION: Epigenetic landscapes in the regulatory regions reflect binding condition of transcription factors and their co-factors. Identifying epigenetic condition and its variation is important in understanding condition-specific gene regulation. Computational approaches to explore complex multi-dimensional landscapes are needed.RESULTS: To study epigenomic condition for gene regulation, we developed a method, AWNFR, to classify epigenomic landscapes based on the detected epigenomic landscapes. Assuming mixture of Gaussians for a nucleosome, the proposed method captures the shape of histone modification and identifies potential regulatory regions in the wavelet domain. For accuracy estimation as well as enhanced computational speed, we developed a novel algorithm based on down-sampling operation and footprint in wavelet. We showed the algorithmic advantages of AWNFR using the simulated data. AWNFR identified regulatory regions more effectively and accurately than the previous approaches with the epigenome data in mouse embryonic stem cells and human lung fibroblast cells (IMR90). Based on the detected epigenomic landscapes, AWNFR classified epigenomic status and studied epigenomic codes. We studied co-occurring histone marks and showed that AWNFR captures the epigenomic variation across time.AVAILABILITY AND IMPLEMENTATION: The source code and supplemental document of AWNFR are available at http://wonk.med.upenn.edu/AWNFR.
AB - MOTIVATION: Epigenetic landscapes in the regulatory regions reflect binding condition of transcription factors and their co-factors. Identifying epigenetic condition and its variation is important in understanding condition-specific gene regulation. Computational approaches to explore complex multi-dimensional landscapes are needed.RESULTS: To study epigenomic condition for gene regulation, we developed a method, AWNFR, to classify epigenomic landscapes based on the detected epigenomic landscapes. Assuming mixture of Gaussians for a nucleosome, the proposed method captures the shape of histone modification and identifies potential regulatory regions in the wavelet domain. For accuracy estimation as well as enhanced computational speed, we developed a novel algorithm based on down-sampling operation and footprint in wavelet. We showed the algorithmic advantages of AWNFR using the simulated data. AWNFR identified regulatory regions more effectively and accurately than the previous approaches with the epigenome data in mouse embryonic stem cells and human lung fibroblast cells (IMR90). Based on the detected epigenomic landscapes, AWNFR classified epigenomic status and studied epigenomic codes. We studied co-occurring histone marks and showed that AWNFR captures the epigenomic variation across time.AVAILABILITY AND IMPLEMENTATION: The source code and supplemental document of AWNFR are available at http://wonk.med.upenn.edu/AWNFR.
KW - Algorithms
KW - Animals
KW - Cell Line
KW - Cluster Analysis
KW - Embryonic Stem Cells/metabolism
KW - Epigenomics/methods
KW - Fibroblasts/metabolism
KW - Gene Expression Regulation
KW - Histones/chemistry
KW - Humans
KW - Lung/metabolism
KW - Mice
KW - Software Design
U2 - 10.1093/bioinformatics/btt467
DO - 10.1093/bioinformatics/btt467
M3 - Journal article
C2 - 24096080
VL - 30
SP - 908
EP - 914
JO - Bioinformatics (Online)
JF - Bioinformatics (Online)
SN - 1367-4811
IS - 7
ER -
ID: 199332127