PeakSeg: constrained optimal segmentation and supervised penalty learning for peak detection in count data

Toby Hocking, Guillem Rigaill, Guillaume Bourque
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:324-332, 2015.

Abstract

Peak detection is a central problem in genomic data analysis, and current algorithms for this task are unsupervised and mostly effective for a single data type and pattern (e.g. H3K4me3 data with a sharp peak pattern). We propose PeakSeg, a new constrained maximum likelihood segmentation model for peak detection with an efficient inference algorithm: constrained dynamic programming. We investigate unsupervised and supervised learning of penalties for the critical model selection problem. We show that the supervised method has state-of-the-art peak detection across all data sets in a benchmark that includes both sharp H3K4me3 and broad H3K36me3 patterns.

Cite this Paper

BibTeX
@InProceedings{pmlr-v37-hocking15, title = {PeakSeg: constrained optimal segmentation and supervised penalty learning for peak detection in count data}, author = {Hocking, Toby and Rigaill, Guillem and Bourque, Guillaume}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {324--332}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/hocking15.pdf}, url = {https://proceedings.mlr.press/v37/hocking15.html}, abstract = {Peak detection is a central problem in genomic data analysis, and current algorithms for this task are unsupervised and mostly effective for a single data type and pattern (e.g. H3K4me3 data with a sharp peak pattern). We propose PeakSeg, a new constrained maximum likelihood segmentation model for peak detection with an efficient inference algorithm: constrained dynamic programming. We investigate unsupervised and supervised learning of penalties for the critical model selection problem. We show that the supervised method has state-of-the-art peak detection across all data sets in a benchmark that includes both sharp H3K4me3 and broad H3K36me3 patterns.} }
Endnote
%0 Conference Paper %T PeakSeg: constrained optimal segmentation and supervised penalty learning for peak detection in count data %A Toby Hocking %A Guillem Rigaill %A Guillaume Bourque %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-hocking15 %I PMLR %P 324--332 %U https://proceedings.mlr.press/v37/hocking15.html %V 37 %X Peak detection is a central problem in genomic data analysis, and current algorithms for this task are unsupervised and mostly effective for a single data type and pattern (e.g. H3K4me3 data with a sharp peak pattern). We propose PeakSeg, a new constrained maximum likelihood segmentation model for peak detection with an efficient inference algorithm: constrained dynamic programming. We investigate unsupervised and supervised learning of penalties for the critical model selection problem. We show that the supervised method has state-of-the-art peak detection across all data sets in a benchmark that includes both sharp H3K4me3 and broad H3K36me3 patterns.
RIS
TY - CPAPER TI - PeakSeg: constrained optimal segmentation and supervised penalty learning for peak detection in count data AU - Toby Hocking AU - Guillem Rigaill AU - Guillaume Bourque BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-hocking15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 324 EP - 332 L1 - http://proceedings.mlr.press/v37/hocking15.pdf UR - https://proceedings.mlr.press/v37/hocking15.html AB - Peak detection is a central problem in genomic data analysis, and current algorithms for this task are unsupervised and mostly effective for a single data type and pattern (e.g. H3K4me3 data with a sharp peak pattern). We propose PeakSeg, a new constrained maximum likelihood segmentation model for peak detection with an efficient inference algorithm: constrained dynamic programming. We investigate unsupervised and supervised learning of penalties for the critical model selection problem. We show that the supervised method has state-of-the-art peak detection across all data sets in a benchmark that includes both sharp H3K4me3 and broad H3K36me3 patterns. ER -
APA
Hocking, T., Rigaill, G. & Bourque, G.. (2015). PeakSeg: constrained optimal segmentation and supervised penalty learning for peak detection in count data. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:324-332 Available from https://proceedings.mlr.press/v37/hocking15.html.

Related Material