Skip to content

Define custom patterns

Casper Steinmann edited this page May 17, 2016 · 1 revision

If you want to create your own patterns for fragmentation and use those custom patterns instead you should read along for an example or two.

Protein Backbone

Initially, FragIt was made for the FMO method in GAMESS. Here, it is customary to cut next to peptide bonds because it behaves like a conjugated bond. It is also not recommended when using the FMO method to cut the disulfide bond in a protein. Thus, the default SMARTS fragmentation pattern for a peptide polymer is

[$(CN)][$(C(=O)NCC(=O))]

which is specified in the [fragmentpatterns] section of the configuration file labelled peptide.

If, however, your plan is to use the MFCC style fragmentation (available in the development snapshots) instead then you might prefer cutting in a different place. Because one caps each fragment to an arbitrary order in FragIt, it makes somewhat more sense to cut the protein at the peptide bond and then let the capping do the rest. A very specific MFCC-style pattern developed by Jógvan Magnus Olsen from the University of Southern Denmark as a replacement for the pattern shown above is

peptide = [$([CX3](=[OX1])[NX3][CX4])][$([NX3][CX3][CX4])]

Do note that this pattern should replace the first pattern shown on this page. However, as you can see from the [fragmentpatterns] section, it would not be a problem having both but the results might be a bit undesireable.

Disulfide Bonds

If you want to fragment disulfide bonds, the following pattern will deal with it

disulfide = [$(SC)][$(SC)]

Developing Custom Patterns

FragIt uses SMARTS to fragment molecules and the FragIt paper details the choice in style for the SMARTS patterns.

The suggested approach is to take your molecule, load it into Avogadro and use the search by SMARTS tool.