Skip to main content ↑ ↓
Topic: MS-DIAL beginner (Read 3278 times) previous topic - next topic

MS-DIAL beginner

Hello everyone,
I'm just approaching to use MS-DIAL. I would like to ask some questions about it.

1) Is it possible to introduce METLIN library in the MS-DIAL searching? If not, do you recommend to use just the libraries downloadable from MS-DIAL website for the human metabolome?
2) Which file should I export for the metabolites quantification and which table column should I consider as a normalized value to quantify metabolites?
3) Is it possible to export only the validated metabolites after the post-processing analysis to discard the false positives?

Thank you to everyone who will reply  :)
Ilaria

Re: MS-DIAL beginner

Reply #1 –
Hello,

1. No- cannot use Metlin unless you buy it (expensive!) and use locally- so possible if you spend USD 5K. Use MSDIAL libraries and also, you can add more LC-MS/MS libraries from MoNA (https://mona.fiehnlab.ucdavis.edu/downloads) and GNPS (https://gnps.ucsd.edu/ProteoSAFe/libraries.jsp). For combining libraries, you can use a simple method like https://www.protocols.io/view/steps-for-building-an-open-source-ei-ms-mass-spect-eq2ly33rqgx9/v1 but for your LC-MS/MS use case use ESI-MS or LC-MS or MS/MS libraries only. Add and combine as many as possible and does not matter Pos or Neg mode as they all help and do not mess up with matching.

2. Export "Area" or "Height" and both works fine for me. If normalized(if you ran multiple pooled QCs)  then hit normalize using LOWESS and QC or using internal standards and then export the "normalized data file" from Alignment Results.

3. Not possible. Export everything and then "filter out your validated". What is your definition of validated? If MS/MS matched with library without RT match, then fine too and can be filtered from the final area/ height file export as "MS/MS pos match" and a "cosine/ dot product" similarity score of say >70 or 0.7.

Hope this helps.

Thanks,
Biswa

Re: MS-DIAL beginner

Reply #2 –
Hello Biswa,
thanks for your reply!
About the point 2, I am trying to normalize with LOWESS or LOWESS+internal standard but the "Normalize" button is not clickable. I set QC in correspondence of the QC run in the table where you specify the type of the data. Do you know why?
Regarding point 3, with the term "validated" I meant the assignment of the true positives compounds. Thank for the suggestion to how find them in the final table!

Re: MS-DIAL beginner

Reply #3 –
Welcome.
You would need:
 1. at least 3 QCs, and then the "clickable" options will come-or else it would not.
2. True Positives also needs a definition from an user than a software tool:
Only MS1 match is not enough to believe the hits!)!
Above with MS2 (MS/MS) against public spectral library at say >70% cosime match is higher confidence (likely publishable).
Above both with RT (if you have an internally generated LC library for your compounds of interest!) then it's of even higher confidence.
If you have IOn Mobility generated CCS values to match then it adds to the confidence.

So you got to take your call!

Thanks,
Biswa