How can I define which samples are controls and which samples are cases in ChAMP
0
0
Entering edit mode
@yuabrahamliu-17670
Last seen 3.7 years ago

Hi there,

I have a question about ChAMP again. After I used champ.DMP to find the differential methylated probes successfully. I deemed the probes with a minus logFC as hypo ones (their methylation level was lower in cases than controls), while that with a plus logFC as hyper ones  (their methylation level was higher in cases than controls). However, when I checked them manually, I found actually, the DMPs with a minus logFC had a lower level in controls, not in cases, which was totally contrary to what was expected by me, so was it mean that I reversed the sample labels when using champ.DMP? Actually I didn't find a parameter here used to assign which samples belong to the case group and which ones belong to the control group. This information was only contained in the SampleSheet.csv file when I loaded the data at the beginning. Hence, I am wondering is it necessary to define the sample group by myself in the champ.DMP step, so that I could guarantee that the control group samples were used as references, not the case group as references? and how can I do that? Thank you so much.

Best wishes,

Yu

 

ChAMP champ.DMP Cases Controls references • 1.1k views
ADD COMMENT
0
Entering edit mode

Hello Yu:

champ.DMP() can only assign "comparegroup", which means you tell program which two phenotypes you want to compare, then the program would automatically ORDER them by their name. So sorry that champ.DMP() can not assign case/control.

However, in the analysis view, the result of DMP should be exactly the same despite of the directions. The DMPs you find via comparing case to control, should also be DMPs you find by comparing control to case. 

You just need to manually add negative sign on the logFC/t statistic, then it should be correct.

By the way, you may also send email to champ450K@gmail.com for ChAMP questions in the future.

Best

Yuan Tian

ADD REPLY
0
Entering edit mode

I see. Thank you Yuan!

ADD REPLY

Login before adding your answer.

Traffic: 812 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6