Simulating Tandem Mass Spectra for Small Molecules using a General-Purpose Large-Language Model

Files for testing the hypothesis that a Google Gemini Model can predict mass spectra for small molecules. 

Gemini_Prediction_Positive_Mass_Spectra_11.04.2025.msp  -- Gemini Predicted spectra (M+H) for 1436 compounds 
Gemini_prediction_Negative_Mass_Spectra_11.04.2025.msp -- Gemini Predicted spectra (M+H) for 835 compounds  
IDSL.CSA_parameters_for_negative_mode.xlsx -- spectra similarity search parameters for negative mode
IDSL.CSA_parameters_for_positive_mode.xlsx  -- spectra similartiy search parameters for positive mode
Negative_mode_spectra_similarity_results.csv  --- results of spectra similarity search (negative mode)
Positive_mode_spectra_similarity_results.csv --- results of spectra similarity search (positive mode)
gemini_prompt_negative_mode.txt ------- chain-of-through prompt for predicting M-H spectra
gemini_prompt_positive_mode.txt ------- chain-of-through prompt for predicting M+H spectra
gemini_spectra_generate.py ------- python script to call the Gemini API and populate the spectral library
negative_mode_nodes_cytoscape.txt --- Cytoscape network visualization node details (M+H)
positive_mode_nodes_cytoscape.txt --- Cytoscape network visualization node details (M-H)
sample-gemini-key.txt  --- file to place the Gemini API key
test_candidate_pubchem.json --- Chemical details obtained from the PubChem database