The present invention relates to proteomics, and techniques for predicting of mass spectrometry data of chains of amino acids, such as peptides, proteins, or combinations thereof. Particularly, aspects of the present invention are directed to a computer implemented method that includes obtaining a digital representation of a peptide sequence, the digital representation including a plurality of container elements, each container element of the plurality of container elements representing an amino acid residue; encoding, using a bidirectional recurrent neural network of long short term memory cells, each container element as an encoded vector; and decoding, using a fully-connected network, each of the encoded vectors into a theoretical output spectrum. The theoretical output spectra are represented as a one-dimensional data set or a multi-dimensional data set including intensity values for each fragment ion including one or more of the amino acid residues in the theoretical output spectra.