abnormal normalized & unnormalized wig files

[Mingyang Cai]

Hi Anton,

I am doing ChIP-seq data analysis using QuEST. The data is from mouse, so the reference I use is mm9. But I think the result I get has some problems. I attach the IGV visualization below. I checked the control (Input) file and find something very weird. 

The first line (KO_K4Me3_control_chr1) and second line (WT_K4Me3_control_chr1) show the distribution of chr1 INPUT for KO and WT, which are the output files of MACS. Clearly, as INPUT, the distribution seems normal. 

The third to sixth lines are output files of QuEST. I learn from Google group that you tend to use the normalized wig files. However, it seems that KO_uH2A_2nd_stringent_background_normalized_chr1, as INPUT, the distribution is not normal; while WT_uH2A_2nd_stringent_background_normalized_chr1 's distribution seems more normal. (I am sorry that I should use K4Me3 but not uH2A to better compare the results parallel.) A senior lab member uses different parameters and he gets this similar trend. He says he never encounter this problem before. So it's very strange to me. Can you provide any insights on this problem? Your help is truly appreciated. 

Best,

Mingyang

[Anton]

Hi Mingyang,

The third to sixth lines are output files of QuEST. I learn from Google group that you tend to use the normalized wig files. However, it seems that KO_uH2A_2nd_stringent_background_normalized_chr1, as INPUT, the distribution is not normal;

I am not sure what you mean by 'not normal'. Are you having difficulty calling peaks? 

QuEST by default reports values to wig files only at positions where the normalized scores exceed 10. This is probably a reason for the 'fragmented' data display

This can be changed in the configuration file 

parameters/background_normalized.wig.batch.pars

by modifying

profile_threshold = 10

you can then edit parameters/QuEST.batch.pars to only generate wig files

# ** Pipeline steps, use # to comment out **

#quick_window_scan: ChIP

#calibrate_peak_shift: ChIP

#generate_profile: ChIP

#generate_profile: background

#report_bad_control_regions: true

#peak_caller: ChIP vs background

#metrics_ChIP: true

#bed_graph_tracks_ChIP: true

#bed_tracks_ChIP: true

#bed_tracks_RX_noIP: true

wig_tracks_ChIP: true

wig_tracks_background: true

#filter_ChIP_calls: true

#calls_tracks: true

[Mingyang Cai]

Hi Anton, 

Thanks for your reply. Sorry for my bad description. When I say "normal", it means that the coverage in wig files for the control in ChIP-seq experiment should be very comprehensive across the whole genome, instead of fragmented spikes far separated. I think you have got my meaning. But I am confused that the two files which you suggest me to edit, background_normalized.wig.batch.pars and QuEST.batch.pars are the output files after I run the QuEST. Does it make more sense to edit some files before running QuEST？Maybe I misunderstand your point somewhere, so could you please tell me more about that? Thanks a lot. 

Best,

Mingyang

--

 
---
You received this message because you are subscribed to the Google Groups "QuEST ChIP-seq group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chipseq+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Anton]

Yes, you can exit after configuration script and edit those files. Then run

run_QuEST_with_param_file.pl -ap <QuEST_analysis>

Anton