All datasets in MassNet are processed through a standardized computational pipeline powered by MSDT-Converter , generating unified AI-ready formats with integrated quality control and provenance tracking to ensure consistency, reproducibility, and reusability for downstream machine learning and AI applications.
2Search and processing details
Search engines
FragPipe (v21.1), Sage (v0.14.6)
Enzyme
Trypsin
Missed cleavages
2
Precursor mass tolerance
20 ppm
Fragment mass tolerance
20 ppm
Fixed modifications
Carbamidomethyl (C)
Variable modifications
Oxidation (M), Acetyl (Protein N-term)
FDR threshold
1% (PSM, Peptide, Protein)
Digest min length
7
Digest max length
50
3AI-ready outputs
Raw spectra
Centroided MS/MS spectra with precursor and scan-level metadata (rawspectrum.parquet)
Parquet
MSDT-converted PSM table
Standardized AI-ready PSM representations generated from Sage and FragPipe search results (sage_msdt.parquet, fp_msdt.parquet)
Parquet
Search results
Database search outputs generated by Sage and FragPipe workflows (sage_search_result, fragpipe_search_result)
