Empirical approach for predicting mass spectra

  • Thread starter Spathi
  • Start date
In summary, the study presents an empirical approach for predicting mass spectra, leveraging machine learning techniques to analyze and model the relationships between molecular structures and their corresponding mass spectral data. By utilizing a large dataset of known compounds, the method aims to enhance the accuracy of mass spectrometry predictions, facilitating quicker identification and characterization of unknown samples in various fields, including chemistry and biochemistry. The approach demonstrates significant potential for improving analytical workflows and advancing research in mass spectrometry.
  • #1
Spathi
Gold Member
101
10
TL;DR Summary
I know about the McLafferty rearrangement and the migration of the double bond; maybe there are some other such well-known rearrangements?
One of ways of predicting a mass spectrum of a new molecule is to run a semi-empirical BOMD computation where the molecule is ionized, gets a big energy and starts dissociating. However, I suppose that in many cases a better approach is using some empirical knowledge about the molecule.
I plan to implement a program where you can specify the structure of your molecule, then type the position of a peak in your mass spectrum, and the program will enumerate all possible fragment ions which can produce this peak. One approach is to exclude all rearrangement ions. However, I suppose that the rearrangement ions occur in the mass spectrums quite oftenly. What are the most typical types of rearrangements in the mass spectrum? I know about the McLafferty rearrangement and the migration of the double bond; maybe there are some other such well-known rearrangements?
I am thinking of implementing the following algorithm for checking possible ions. Consider you have this molecule:

1719117849229.png

The program will enumerate all pairs of bonds, remove these bonds and see possible structures obtained via the re-unition of the fragments. For example, when it removes the C1-C11 and C4-C21 bonds, these recombination fragments will be possible:

1719118027807.png

1719118048104.png

1719118081037.png

Would such approach be useful for chemists dealing with mass spectrometry?
 
Last edited:
Back
Top