Keyerror: normalizeBedgraph.py

[Prathyusha-konda]

Hello,

I am getting the following error while executing the normalizebedgraph file.

At line 49 in
if start > chrToLimit[row[0]] or stop > chrToLimit[row[0]]: continue
Keyerror: '20'

Please help me with this.

Thank you..

[zhenyisong]

@Prathyusha-konda I also tried to replicate their results published in Mol. Cell, 2011 to QC my own pipeline <blood_pipeline_QC.sh>.
Here is the link to their own protocol by Qu Kun, the co-author in the above paper.
Speaking of your python running error, did you check the python version? Py2 and Py3?

I still have my own problem of replicating their roX2 data. Anyway, hope you can find the solution to your confusion. Z

[Prathyusha-konda]

Hi @zhenyisong, Thanks for your reply. I figured out that the error in my case was from the headers of the sizes file. They were different as compared to the ref file. it worked for me after fixing that.

I am trying to do the analysis on my data similar to what they have done in their Mol Cell paper. I used their perl script from the website as a reference and modified the parameters and code accordingly. It seems to have worked fine for me.

May be you should try the perl script too. I hope it works for you.

[zhenyisong]

@Prathyusha-konda Thanks. I have completed the data analysis replication published in the Mol. Cell paper and confirmed the specificity of the binding pattern by the rox2 using their fly data set. I think they used the filtering parameters to control the specificity ( or the true targets). You can read my two scripts: blood_pipeline_QC.sh and blood_pipeline_QC.R.

The Perl scripts at their site (Stanford) seems to only accept SE library model. If the results are paired-end sequencing model. the results are perhaps not what we expect. I chose another way.

BTW, do you design any protocol to confirm the specificity of RNA-DNA binding patterning? Any filtering parameter consideration? In their old paper, the rox2 binded most of the sex chromosome regions.