|Alternative name||Zero-Inflated Negative Binomial Algorithm|
|Interface||Command line interface|
|Restrictions to use||None|
|Operating system||Unix/Linux, Mac OS, Windows|
|License||GNU General Public License version 2.0|
Add your version
- person_outline Jason Lieb <>
- person_outline Wei Sun <>
Publication for Zero-Inflated Negative Binomial Algorithm
[…] reference, were removed., read alignment positions were adjusted according to their strand: +4 bp for + strand alignments, and −5 bp for − strand alignments. open chromatin regions were called using zinba53 with a window size of 300 bp, an offset of 75 bp, and a posterior probability threshold of 0.8., for nucleosome positioning, properly paired alignments were filtered by their fragment size. […]
[…] was allowed to occur up to four times throughout the genome, where one was selected at random for each of those positions that were not unique. the set of enriched regions were then identified by zinba, using 300-bp windows with 75-bp offsets [see additional files 2 and 3]. the background and enriched components were modeled using g/c content and an interaction term between mapability […]
To balance my comment, I have to say that I used ZINBA on FAIRE-seq which is relatively more noisy compared to ChIP-seq.
But another black point is that there is almost no support for a lot issues and instabilities.
For FAIRE-seq I would prefer Fseq and for ChIP-seq I would prefer MACS.