MS2PIP

MS2PIP predicts peptide fragmentation MS/MS spectra using machine learning to generate predicted spectral libraries for proteomics applications such as immunopeptidomics, data-independent acquisition (DIA) full-proteome identification, and TMT-based quantification.

Key Features:

• Enhanced prediction models: Updated machine learning models improve performance for tryptic and non-tryptic peptides, immunopeptides, and collision-induced dissociation (CID)-fragmented TMT-labeled peptides.
• Retention time prediction: Integrates DeepLC for retention time predictions to complement fragment intensity predictions.
• Spectral library generation: Generates proteome-wide predicted spectral libraries from FASTA protein files and includes pre-built spectral libraries for multiple model organisms in several DIA-compatible formats.
• Support for TMT/MS3 experiments: Extended functionality to handle TMT-labeled peptides and applicability to MS3-based TMT quantification experiments.
• Input format: Accepts FASTA protein files for proteome-wide prediction workflows.

Scientific Applications:

• Immunopeptidomics: Facilitates identification and characterization of peptides presented by major histocompatibility complex (MHC) molecules.
• Full-proteome DIA identification: Enhances analysis of DIA spectra to support comprehensive proteome coverage and spectral-library–based identification.
• TMT quantification experiments: Supports quantitative proteomics workflows involving TMT labeling and MS3-based quantification.

Methodology:

Uses machine learning-based prediction models for tryptic and non-tryptic peptides, immunopeptides, and CID-fragmented TMT-labeled peptides, applies DeepLC for retention time prediction, and generates predicted spectral libraries from FASTA protein files.