Improving protein order-disorder classification using charge-hydropathy plots

Area of Science:

• Protein bioinformatics
• Computational biology
• Biophysics

Background:

• The earliest protein disorder predictor, the charge-hydropathy (C-H) plot, used the Kyte-Doolittle hydropathy scale.
• The performance of the C-H plot in distinguishing structured from intrinsically disordered proteins (IDPs) may be improved by alternative scales.

Purpose of the Study:

• To evaluate alternative hydropathy scales for improving the C-H plot's predictive performance.
• To develop a novel hydropathy scale optimized for IDP prediction.

Main Methods:

• Compared 19 hydropathy scales using the C-H plot's classification accuracy.
• Developed and tested a new scale, IDP-Hydropathy, on a dataset of structured proteins and IDPs.

Main Results:

• The Guy hydropathy scale improved C-H plot performance over the original Kyte-Doolittle scale.
• The novel IDP-Hydropathy scale achieved 90% accuracy in classifying structured proteins and IDPs, surpassing existing scales.
• IDP-Hydropathy showed strong correlation with other scales, indicating specificity to hydropathy.

Conclusions:

• The IDP-Hydropathy scale offers superior performance for predicting protein disorder.
• This new scale is recommended for algorithms aimed at disorder prediction.