Predicting antigenic determinants in proteins: looking for unidimensional solutions to a three-dimensional problem?

Area of Science:

• Immunology
• Computational Biology
• Protein Science

Background:

• The Hopp and Woods hydrophilicity method is widely used for predicting B cell epitopes.
• Investigators aim to produce peptide-protein cross-reactive antisera using antigenicity prediction algorithms.
• Current prediction methods rely on amino acid propensity scales for properties like hydrophilicity.

Purpose of the Study:

• To challenge the asserted superiority of the Hopp and Woods hydrophilicity method.
• To evaluate the efficacy of various antigenicity prediction scales.
• To discuss reasons for the low success rate in predicting B cell epitopes.

Main Methods:

• Comparison of the prediction efficacy of 22 different amino acid propensity scales.
• Assessment of prediction accuracy considering both correct and incorrect predictions.
• Analysis of the methodology used to evaluate prediction success rates.

Main Results:

• No tested scale achieved a correct prediction rate higher than 50%-60%.
• Hydrophilicity peaks in regions lacking antigenicity were considered incorrect predictions.
• The reported high success rate of the Hopp method is attributed to its specific assessment criteria.

Conclusions:

• The Hopp and Woods hydrophilicity method's efficacy is overestimated.
• Current antigenicity prediction methods have limited success rates.
• Linear peptide models are often inadequate for complex, discontinuous epitopes.