Comparing big data sets needs normalization. I have lipid data from tissues and entire animals.
We add an internal calibration standard that corrects for technical parameter, like pipetting.
I align and pick the peaks with the QI software from nonlinear.
What are your suggestions?
What I have been getting from seminars, experience, talks, and from individuals:
1. People use isotopic lipid for each class of lipids they targeted for analysis/ quantification. This is a huge task to collect and use 17-20 internal standards for each class and do the normalization class-wise and so on.
2. 2-8 internal standards either isotopic or non-isotopic that is consistent and dependable also does the job. But most studies 5 years back depended on just 1 internal standard and did not have to be a lipid !
3.If nothing was used prior to extraction and mass-spec runs, still "sum" normalization or "median normalization" would work great. The MIA-Metabolomics-Experiments are still on the way / "lacking" as far as all classes of compounds, platforms etc. are to be considered !
Just my 2 cents based on reading and interacting (not doing lipidomics)! best wishes. Biswa
If you have a spiked internal standard and known amount of tissue/cell count before the mass spec analysis was performed, you can normalized based on the peak area. During my graduate training we did this as a standard protocol for LIPIDMAPS and our regular studies. Here are a few references. By the way do you have lipid standards for each lipid family or species ?
1) A method for visualization of "omic" datasets for sphingolipid metabolism to predict potentially interesting differences. J Lipid Res. 2011 Jun;52(6):1073-83. doi: 10.1194/jlr.M010454. Epub 2011 Mar 17.
2) Sphingolipidomics: high-throughput, structure-specific, and quantitative analysis of sphingolipids by liquid chromatography tandem mass spectrometry
AH Merrill, MC Sullards, JC Allegood, S Kelly, E Wang
Methods 36 (2), 207-224
Spiked in standards are acceptable, however, it is important that they be in the range (with respect to concentration, m/z, and retention time) as the analytes you are interested in measuring. Also- beware of suppression.
Here is a reference that has good summaries this method and others:
http://www.ncbi.nlm.nih.gov/pubmed/17362505
http://www.ncbi.nlm.nih.gov/pubmed/17362505
- Badges
- Science method