Metabolomics Society Forum

Software => R => XCMS => Topic started by: yashwant on March 10, 2013, 11:39:35 PM

Title: time series experiment
Post by: yashwant on March 10, 2013, 11:39:35 PM
I have acquired data of pathogen inoculated susceptible and resistant cell-lines along with their controls (mock inoculated) at each time point. I have time points of 2 hr, 4hr, 8 hrs, 16 hr. I request you to let me know how to group the data.
I have tried following approach, please let me know whether I am correct or not?
First I have made a folder control and infected, within these folder I have made a subfolder of 2 hrs, 4hr, 8hr and 12 hr.
Title: Re: time series experiment
Post by: hpbenton on March 15, 2013, 11:45:56 AM
yashwant,

It depends on what you want  in some ways. Really the main thing here is if you use the density grouping method with the minFrac at 50% you will lose some peaks. I would think that setting up the groups by time points would be fine. I have in the past also just put everything into class folders and then separated everything out afterwards. This way I have the t-test for the classes. It can be helpful to tell me if something didn't change particularly through time but rather just between classes.

Just reread your question. :) If I remember correctly you're going to have some problems with retention time alignment if you use subfolders.

Does that all make sense. Let me know how you get on.

Cheers,

Paul