Skip to main content ↑ ↓
Topic: alignment accross batches (Read 5009 times) previous topic - next topic

alignment accross batches

Hi everyone,
I use xcms for data processing obtained from LC-QTOF-MS. We analyzed samples in multiple batches due to the large amount of sample size. However, it seems that the batch effect on retention time is very severe. Specifically, I use the obiwarp funtion for retention time correction. If I seperate my XCMS object into differnt batches and do the retention time correction for each batch,  based on the retention time deviation plot, I should set the bw is about 2 for grouping (retention time deviation is below 2). However, if I do the retention time correction for whole XCMS object (8 batches), the retention time deviation is over 15!
Moreover, as you can see from the plot, according to the rmsd, as calculated as
 rmsd<-sapply(1:length(hilicNEG_XS_re@rt$corrected),function(x)
{sqrt(sum((hilicNEG_XS_re@rt$raw[
  • ]-hilicNEG_XS_re@rt$corrected[
  • ])**2)/length(hilicNEG_XS_re@rt$corrected[
  • ]))
})
the difference between batches is very obvious.
So in this case, I am wondering, whther I use bw=15 as indicated as the retention time deviation plot accross whole batches, or bw =3 as indicated as  the retention time deviation plot for one batch, features would be miss aligned!!
How can I deal with this sitiuation??
Thank you in advance for valuable suggestions !!!!

[attachment deleted by admin]