Can I run ENCODE ChIP-Seq pipeline or get their code?

Question

[rd]

I found some interesting ENCODE CTCF ChIP-Seq samples that I would like to compare to some other publicly available data. To avoid any technical artifacts caused by the difference in pipelines used, I would like to process all the publicly available samples using the ENCODE ChIP-Seq pipeline to be able to compare everything to the ENCODE CTCF ChIP-Seq samples that I am interested in. Is it possible to find the code/pipeline ENCODE uses? Have they made it publicly available? I spent a few hours on google but that hasn't resulted in anything yet.

I have found general pipeline notes on their github account like:

"1- Map reads with BWA, mark duplicates Picard, and remove duplicates. 2- Estimate library complexity and calculate calculate NRF (non-redundant fraction), PBC1, PBC2 (PCR bottleneck coefficient). 3- Calculate cross-correlation analysis with spp/phantompeakqualtools. 4- Generate p-value and fold-over-control signal tracks for each replicate and replicates pooled with MACS2. Call peaks with MACS2."

Every tool has a set of parameters that can alter the outcome a little bit. So I wanted to check whether the actual pipeline is available before I write a pipeline based on the general description I found in papers.

Thanks!

Answer 1

[Santosh Anand]

http://encodedcc.stanford.edu/encodevm/

[rd]

I downloaded the ENCODEVM and looked through most folders. There are Rscripts to make their plots but no ChIP-Seq pipeline. Most scripts seem to take as input already processed ChIP-Seq data. Did I miss something?

[Santosh Anand]

Hmmm. Check if their IDR framework has got the thing that you are looking for https://sites.google.com/site/anshulkundaje/projects/idr

Answer 2

[karl.sebby]

The RNA-seq pipelines have a list of steps and associated bash scripts that show how the tools are being used (https://github.com/ENCODE-DCC/long-rna-seq-pipeline/tree/master/dnanexus), but I don't see that for ChIP-seq pipeline. I have found it useful to look at experiments that use the pipeline, rather than just the pipeline itself, e.g. https://www.encodeproject.org/experiments/ENCSR670JDQ/.

[karl.sebby]

The easiest solution would be to run it on DNAnexus platform. Or you could always try to get an account and see if you can see parameters there. I tried to get an account with my gmail address but was denied-- it wanted an institutional address so I'm not sure how hard it is to get onto their platform just to look around.

[rd]

I also noticed that some of the ENCODE pipelines have explicit code on github which is why I thought I might be missing something for ChIPSeq.

I made an account on DNAnexus and it seems to have workflows one can use on the platform but I can't seem to extract parameters to run on my machine. In some ways it resembles Galaxy. If that is my only option then I will analyze the data through the DNAnexus portal. Right now i am downloading The ENCODE virtual machine at the suggestion of Santosh Anand. I will report back if I find it useful.

Answer 3

[GouthamAtla]

You can definitely run. All the scripts are on github with a detailed protocol.

https://github.com/kundajelab/chipseq_pipeline

[rd]

Thanks! I will try to go through it.