dpos ChiP-seq input, spike-in, replicates

[Gilles Siréta]

Hi all,
I have three questions about the use of DANPOS' dpos function, and answers to even just one of them would be greatly appreciated.
As a preface, I am using DANPOS on H3 ChIP-seq data (in Arabidopsis thaliana).

1. About the -b (--bg) parameter : can this be used to provide the ChIP-seq input (by which I mean the results for a sample just like the IP, but which did not undergo immunoprecipitation, not a mock IgG) ? The description seems to imply this is not really what it is used for, but leaves (for me) some doubts, and any further information would be appreciated.
2. Can/Should spike-in normalization be done through DANPOS itself ? I have found the  "-c , --count" parameter but I am uncertain as to its use, as with spike-in a ratio of the DNA belonging to the studied organism vs spike-in organism is typically used, using the input results to properly normalize the IP, but this asks for a reads number.
3. What is the right way to integrate multiple replicates into a command ? I have 2 replicates and would like to provide them both in the same command. As we also have two different conditions to compare, my colleague has settled on this : put both replicates in a single folder, one folder per condition, then use the command like this : "danpos.py dpos Condition1/:Condition2/" Does this seem right to you ?

Many thanks in advance for any assistance you are able to provide,
Gilles Siréta

[simon holzinger]

Hi Gilles, 

First I am not the original developer of the program and although I used it a lot and went through quite a lot of the source code I still do not understand every aspect of it. So here are just some of my thoughts

1. The -bg parameter does the following: The bg input gets smoothed (smoothing window is controlled by -l/-lmd) and then subtracted from all samples. This leads to a kind of "soft" background removal, which means you will still find peaks if you use the -bg with the same data as the input. The parameter has its use but I would be careful. Maybe have a few tries and see if the result is what you want/expect. 

2. The -c parameter can be used for read-count normalisation between samples. But If you know the ratios between your samples from your spike in you can use it to normalise to those ratios as well.

3.  I do it the exact same way, you are planning to do. Just one remark: DANPOS will only pool the replicates and not do any statistics with those.

Bw

Simon