Denoising search: for single spectrum

The denoising_search function, which is the core function of the project, performs identity search with spectral denoising integrated. In this way, the query spectrum is denoised using all molecular information obtained from candidate spectra with predefined precursor mass range. The entropy similarity scores are computed with denoised spectra.

Example usage: The demo data can be found here.

import pandas as pd
import spectral_denoising as sd
quene_spectra= sd.read_msp('sample_data/query_spectra.msp')
reference_library =sd.read_msp('sample_data/reference_library.msp')
quene_spectrum, quene_pmz = quene_spectra.iloc[0]['peaks'], quene_spectra.iloc[0]['precursor_mz']
result = sd.denoising_search(quene_spectrum, quene_pmz, reference_library)

The result will give all candidate spectra within the precursor mass range, with additional column of ‘query_peaks’ (query spectrum), ‘query_peaks_denoised’ (denoised query spectra), ‘entrpy_similarity’ (entropy similarity of query spectra to reference spectra), and ‘denoised_similarity’ (entropy similarity of denoised query spectra to reference spectra).

References

spectral_denoising.denoising_search(msms, pmz, reference_lib, identitiy_search_mass_error=0.01, mass_tolernace=0.005, pmz_col='precursor_mz', smiles_col='smiles', adduct_col='adduct', msms_col='peaks', first_n=1, need_sort=True)