Prediction of LCMSMS properties of peptides from sequence by deep learning

Highlights

Deep learning models for prediction of LCMSMS properties

In Brief Statement

Indexed retention times (iRT), MS1 or survey scan charge state distributions, and sequence ion intensities of MSMS spectra were predicted from peptide sequence by use of long-short term memory (LSTM) recurrent neural networks models.  Data points on order of 105 were used to train the iRT and charge state distribution models.  HCD sequence ion prediction model was trained with 2X106 experimental spectra.  The trained models with a simple deep learning architecture outperform the start-of-the-art models available in literature.