Journal Article

PeaKDEck: a kernel density estimator-based peak calling program for DNaseI-seq data

Michael T. McCarthy and Christopher A. O’Callaghan

in Bioinformatics

Volume 30, issue 9, pages 1302-1304
Published in print May 2014 | ISSN: 1367-4803
Published online January 2014 | e-ISSN: 1460-2059 | DOI:

More Like This

Show all results sharing this subject:

  • Bioinformatics and Computational Biology


Show Summary Details


Summary: Hypersensitivity to DNaseI digestion is a hallmark of open chromatin, and DNaseI-seq allows the genome-wide identification of regions of open chromatin. Interpreting these data is challenging, largely because of inherent variation in signal-to-noise ratio between datasets. We have developed PeaKDEck, a peak calling program that distinguishes signal from noise by randomly sampling read densities and using kernel density estimation to generate a dataset-specific probability distribution of random background signal. PeaKDEck uses this probability distribution to select an appropriate read density threshold for peak calling in each dataset. We benchmark PeaKDEck using published ENCODE DNaseI-seq data and other peak calling programs, and demonstrate superior performance in low signal-to-noise ratio datasets.

Availability and implementation: PeaKDEck is written in standard Perl and runs on any platform with Perl installed. PeaKDEck is also available as a standalone application written in Perl/Tk, which does not require Perl to be installed. Files, including a user guide, can be downloaded at:


Supplementary information: Supplementary data are available at Bioinformatics online.

Journal Article.  1519 words.  Illustrated.

Subjects: Bioinformatics and Computational Biology

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content.