The first model for predicting complete tandem mass spectra from LONG peptides sequences.

Free for academic uses.

• 2024.05.01: First version.

This model follows the main structure of Predfull, but with significant change of the input embedding

• This model assumes a FIXED carbamidomethyl on C
• The maximum input charge is 30+
• The prediction will NOT output peaks with M/z > 2000

Recommend to install dependency via Anaconda

• Python >= 3.9
• Tensorflow >= 2.14.0
• Pandas >= 0.20
• pyteomics
• lxml

Simply run:

python train_model.py --input example.mgf --output model.h5

Note that intensities are shown by square rooted values

CID example:

HCD example: