question about retention time correction

September 26, 2012, 11:29:58 AM

Hi to all,

I have a problem concerning lc-ms data alignment.
Indeed, we have injected a lot of samples in a first batch composed of QC and experimental samples and then we have injected an other batch of samples containing the same QC (i.e. technical replicates), some samples already injected in the first batch and new samples.

The problem is that between batch1 and batch2 the LC was repaired and high retention time deviations appeared between the two batches.

I tried to correct that with the "group" and "retcor" steps but it doesn't work very well.
Firstly I changed the "bw" option in "group" increasing the value but the alignment was not well done : correlation between QC was not good, some peak areas were not integrated (aligned I think) for the second batch QC.
Then, I tried to apply successive rounds of "group" and "retcor" (loess method). At each round, I lowered progressively the "bw" value. Surprisingly in a first attempt 3 rounds were applied successfully and peak areas were well integrated.
As it worked quite well, I tried this strategy on an other study with the same problem and the same kind of experimental design and this time I was only able to apply 1 round of "group"/"retcor" because the second "retcor" induced a shift of 2 or 3 chromatograms to negatives RT while the others chromatograms were still anchored to 0 and aligned each others.

Have you some advices to succeed in aligning chromatograms with high retention time deviations and also to explain this problem ?

Thanks for any help.

Ben

Re: question about retention time correction

Reply #1 – October 08, 2012, 05:05:24 PM

LC was repaired in the middle of the experiment ?
You might not like this answer: for reliable results you should rerun your samples!

in general, the obiwarp method seems to handle larger retention time deviations much better than the peakgroups method.