MGF2MSP connverter breaks SMILES codes

[tobigithub]

The current implementation of the IDSL.FSA mgf2msp function breaks SMILES, probably because the escape characters are not correctly handled. Also for some compounds the SMILES code is completely transformed, which should not be the case for a simple format converter, unless explicitly requested in a function call.

mgf2msp(path = temp_wd, MGFfile_name)

For diverse inputs SMILES the error rate can can be 50% or higher. Example:

MGF:
SMILES=Brc1cc2c(cc1/C=N/c1c(-c3ccco3)nc3ccccn13)OCO2
MSP:
SMILES: Brc1cc2c(cc1/C: N/c1c(-c3ccco3)nc3ccccn13)OCO2

observe the C=N is converted to C: N (with space). Outputs have to be checked via InChIKey.

tester.msp.txt
tester.mgf.txt

More example files are here.

[barupal]

Thanks, Tobias. We will fix it soon.