IGV 2.17 methylation coloring ISSUE

[majnu pahara]

I recently upgraded IGV to the version 2.17 and I encounter an issue with visualization of Pabcbio 5mc methylation. I can only see methylation coloring on part of the reads. 

Previous version used to show colors on all reads. Can anyone suggest what am I doing wrong? 

Thank you!

[igv-help]

IGV 2.17.2 and above do validation of the methylation tags, previous versions did not.   Depending on your pipeline and the exact tools you use methylation tags can be incorrect if the alignments are clipped.   That could be the reason, but I really can't say for sure from screenshots.   To see if this is the case you can turn validation off from the preferences, "View > Preferences > Base Mods" and uncheck "Validate base count..." at the bottom.  Then reload your alignment file and compare.   This is just a check,  we do not recommend you leave validation off. 

The basic validation rules is the number of "C"s in the read sequence recorded for the alignment must be at least equal to the number of "C"s required by the methylation tag.  If you believe an alignment has failed incorrectly right click over it ,  then select "copy read details to clipboard",  and "copy sequence", and paste the results here and I will look at it.   

[majnu pahara]

Thank you very much. Yes, the methylation validation seems to be the reason it looked different. Below is the same view after I remove the methylation validation as you suggested. It solves my problem for now. Thank you for you help.  

Would like to know more behind the science/process of methylation validation, but I can find else where about that if that is not the scope of this group!

 

--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/qESJCoq02U0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/6448080a-1f29-4a09-a60d-efbf01a34031n%40googlegroups.com.

[James Robinson]

Are the alignments that are not showing methylation hard clipped?  If you aren't sure select look at the alignment details, either mouse over or click depending on your settings, and look at the "Cigar" string.  If it has the letter "H" in it the alignment is hard clipped.   

What you were seeing before, or what you see with validation off, is almost certainly wrong for those alignments.   If you want to see methylation for all alignments look for settings on your alignment tool to turn off hard clipping.  For BWA I think the option is "-Y".  Hard clipping can have other side effects, any tag that relies on the full read sequence is almost certainly wrong after hard clipping.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CAOTEwOE_2Bm%2B92uw1CNwc_Ew8DROH5uwZu4OUXnHv%3DsVfcMBLw%40mail.gmail.com.

[James Robinson]

WRT the "science" of methylation validation there's really no science to this, it's just arithmetic.    The methylation tag will imply a certain number of "C"s for 5mC in the sequence (or "G" if a negative strand alignment).   If the read sequence has fewer "C"s than this number validation fails.  The usual, perhaps the only, cause for this is hard clipping.