Your example data contains compound 'Example_3' which does result in a predicted spectra containing mz 311.938 in total 4 times. The reason being that 4 highly similar fragments with slightly different SMILES are predicted. While I get the point that this reflects how FIORA works, the output would require postprocessing before being used for i.e. spectra comparison or other downstream processing steps. Specifically, when comparing the FIORA predicted spectrum against a measured spectrum redundant mz peaks may lead to errors and definitely will lead to wrong results.
I would suggest to include an option to FIORA to force post processing the output within FIORA. The 4 peaks should be combined into a single row and their relative intensities be summed up. To reflect the presence of multiple SMILES, I would include the most abundant SMILES in the respective column/slot and add the information that 3 additional SMILES were predicted into the comment slot.
Your example data contains compound 'Example_3' which does result in a predicted spectra containing mz 311.938 in total 4 times. The reason being that 4 highly similar fragments with slightly different SMILES are predicted. While I get the point that this reflects how FIORA works, the output would require postprocessing before being used for i.e. spectra comparison or other downstream processing steps. Specifically, when comparing the FIORA predicted spectrum against a measured spectrum redundant mz peaks may lead to errors and definitely will lead to wrong results.
I would suggest to include an option to FIORA to force post processing the output within FIORA. The 4 peaks should be combined into a single row and their relative intensities be summed up. To reflect the presence of multiple SMILES, I would include the most abundant SMILES in the respective column/slot and add the information that 3 additional SMILES were predicted into the comment slot.