IPC 2.0

IPC 2.0 predicts the isoelectric point (pI) and residue dissociation constants (pKa) of proteins and peptides from amino-acid sequence to support biochemical and proteomic analyses.

Key Features:

• Predictions: Estimates isoelectric point (pI) and residue pKa values for proteins and peptides from sequence data.
• Machine learning models: Employs a combination of deep learning and support vector regression (SVR) models for prediction.
• Accuracy — proteins: Achieves RMSD of 0.848 for protein pI predictions compared with 0.868 for previous methods.
• Accuracy — peptides: Achieves RMSD of 0.222 for peptide pI predictions compared with 0.405 for previous methods.
• pKa sequence-based accuracy: Predicts pKa from sequence with RMSD of 0.576 versus structure-based methods with RMSD of 0.826.
• Computational efficiency: Sequence-based pKa predictions are reported to be several times faster than structure-based approaches.

Scientific Applications:

• Two-dimensional gel electrophoresis (2D-PAGE): Provides pI estimates to guide protein separation by isoelectric focusing.
• Capillary isoelectric focusing (cIEF): Supplies pI values to optimize focusing conditions.
• X-ray crystallography: Supplies pI and pKa information relevant to crystallization and buffer selection.
• Liquid chromatography–mass spectrometry (LC-MS): Informs separation strategies and interpretation of proteomic LC-MS data.

Methodology:

Uses deep learning and support vector regression models to predict pI and pKa from amino-acid sequence and reports performance using RMSD comparisons against previous and structure-based methods.