Skip to main content ↑ ↓
Topic: Metlin hits (Read 5929 times) previous topic - next topic

Metlin hits

Hi,

I just started playing with some urine samples and try to create an easier workflow between XCMS-online/Metlin/MassTrix.

Some Qs I've collected:
1. Does the metlin compound search in XCMS-online take in account the spectra and any detected isotopes?
2. If 1 is false, then it means that when I search Metlin (I do a 500 compound batch) based on the m/z from diffreport and under the same parameters I'm supposed to find exactly the same hits?
3. Following 2, Wouldn't it be logical to omit any C13 isotopes from my metlin batch search (hence include only [M] and non-isotopic m/z)?
4. I'm importing the results tsv file to excel, and find it time-consuming to manually parse the metlin hits column, which includes 4 fields for each hit (dppm::name::adduct::id ). for example:
 2::2-Acetolactic acid::M-H::  293|2::3-hydroxy-3-methyl-2-oxo-Butyric acid::M-H:: 3253|2::Glutaric acid::M-H:: 3254|...
Is there any way around it apart from learning R so altogether I could handle the files with more flexibility?

Many thanks for anyone who finds the time to respond.

cheers

Re: Metlin hits

Reply #1 –
Quote
1. Does the metlin compound search in XCMS-online take in account the spectra and any detected isotopes?
2. If 1 is false, then it means that when I search Metlin (I do a 500 compound batch) based on the m/z from diffreport and under the same parameters I'm supposed to find exactly the same hits?

The default identification module does not use additional information and you should get the same result if you select the same adducts and ppm.
However, there is another identification module that takes into account the adduct annotation as computed by the CAMERA module. It will automatically search the adduct that was detected in the annotation step, i.e. directly look up the neutral mass. Of course this method only works if full annotation [isotopes + adducts] was selected, which can take longer.

Quote
3. Following 2, Wouldn't it be logical to omit any C13 isotopes from my metlin batch search (hence include only [M] and non-isotopic m/z)?
Not sure what you mean here. XCMS Online always excludes isotopic peaks from the database search.
The batch search on METLIN is not doing additional processing but uses the peak list as provided by the user.

Quote
I'm importing the results tsv file to excel, and find it time-consuming to manually parse the metlin hits column, which includes 4 fields for each hit (dppm::name::adduct::id ). for example:
2::2-Acetolactic acid::M-H:: 293|2::3-hydroxy-3-methyl-2-oxo-Butyric acid::M-H:: 3253|2::Glutaric acid::M-H:: 3254|...


It is a little challenging to represent multiple hits in a single row in a table. The "exotic" separators (|,:::) are used to avoid characters that are part of compound names. The separators can be used by other software to parse the information. Please let me know if you have suggestions for a better representation of multiple hits.

Re: Metlin hits

Reply #2 –
Hi Ralf,

Thanks for your detailed response. it is much more clear now!

1,2. Yes, I used the isotope+adduct option. so correct me if I got it wrong - it means that the isotopes data adds constraints to the compound search and match, but if no isotopes were detected, a (less accurate) match can be found? (many hits on non-isotopic compounds in my list, but less for isotopes [M]).
3. I just mean that I should do the same as the XCMS when I conduct the manual searches on Metlin, and not include m/z of isotopes on my batch search.
4. I was thinking maybe creating 4 different fields (dppm; name; id...), and in each of them have the values separated by |, of course in the same order. so if parsed field by field, each in loop, it would still match according to the order (when inserted to file/db record). For me it will just be easier because I'm interested in the id only ; ) I would totally understand if it doesn't make sense to others.

cheers