Peptide retention time prediction using LSTM artificial neural network architecture

In order to reduce the complexity of peptide-mixture in shotgun proteomics, the liquid chromatography technique has been employed frequently. In these systems, the time when a peptide is released from a chromatography column and registered in the mass spectrometer is referred to as the peptide’s retention time. The reproducibility of this chromatographic separation process [1,2] introduces an interesting prediction problem; predicting a peptide’s retention time by its amino acid sequence. Accurate prediction of retention times can benefit the field of proteomics both by increasing the number of peptide identification and by increasing the reliability of those identification [3]. Moreover, the experimental retention time of the peptide can be compared to the predicted retention time of the identified peptide as an additional verification of the identification [4].