For the record if anybody is interested, just yesterday I started re-writing a wrapper around the SICER pipeline. The package is here https://github.com/dariober/SICERpy. Among other things, I make sicer accept bam files rather then bed.
I started using SICER and it gives sensible results but the user interface is a bit frustrating...!
Great work! Do not work too hard on it though, I have an all parallel numpy/scipy version coming out that uses minutes where SICER used hours (or did not work at all) which is much easier to use ;). Complete rewrite! I have had it lying around for a year, tweaking it. Now that I have spilled the beans I finally am commited to releasing the preprint (will do so within April 14).
Edit: not to diminish your work or anything, just felt bad not letting you know.
Great, looking forward to that and thanks for the heads up! I thought someone MUST have done something... Sicer is still very popular, the encode project used it but it's really messy and inefficient. Keep us posted!
Come to think of it, SICERpy is incredibly valuable. There is loads of functionality that exists in SICER that I do not intend to implement. Thanks for the script!
You'll need to align your fastq files with whatever aligner you prefer (e.g., bowtie2, or bwa) and then convert the resulting SAM or BAM file into a BED file (yes, this seems incredibly silly). Bedtools has a bamtobed command that can facilitate this last step.
That looks like a big step forward, I hope the original authors incorporate this!
Great work! Do not work too hard on it though, I have an all parallel numpy/scipy version coming out that uses minutes where SICER used hours (or did not work at all) which is much easier to use ;). Complete rewrite! I have had it lying around for a year, tweaking it. Now that I have spilled the beans I finally am commited to releasing the preprint (will do so within April 14).
Edit: not to diminish your work or anything, just felt bad not letting you know.
Great, looking forward to that and thanks for the heads up! I thought someone MUST have done something... Sicer is still very popular, the encode project used it but it's really messy and inefficient. Keep us posted!
https://github.com/endrebak/epic or epic: new ChIP-Seq caller based on the SICER algorithm
Please try it and report any issues (I'm the only one who has used it so bound to be some)
Thanks for making that script, I'm not sure whether epic would ever have been released if not for sicerpy.
Cool, thanks man! I'll give it a go...
Yeah, I think I know that feeling I've got this piece of code that looks ok... but who cares?!... Well... some people do!
Come to think of it, SICERpy is incredibly valuable. There is loads of functionality that exists in SICER that I do not intend to implement. Thanks for the script!