This article presents an improved method for predicting linear B-cell epitopes, which are crucial for vaccine design and diagnostic tests. The authors developed BepiPred, a combination method that integrates a hidden Markov model (HMM) with one of the best propensity scale methods (Parker et al.). The HMM was trained using positive windows extracted from the Antijen database, while the propensity scale method was optimized on the Pellequer dataset. The performance of BepiPred was validated using an independent HIV dataset, showing significantly better accuracy compared to other methods tested. The study also compared various propensity scale methods, finding that the scales by Parker et al. and Levitt performed best. The authors conclude that BepiPred is a reliable tool for predicting linear B-cell epitopes and that further improvements are needed to enhance its predictive power. The method and data sets are publicly available for research purposes.This article presents an improved method for predicting linear B-cell epitopes, which are crucial for vaccine design and diagnostic tests. The authors developed BepiPred, a combination method that integrates a hidden Markov model (HMM) with one of the best propensity scale methods (Parker et al.). The HMM was trained using positive windows extracted from the Antijen database, while the propensity scale method was optimized on the Pellequer dataset. The performance of BepiPred was validated using an independent HIV dataset, showing significantly better accuracy compared to other methods tested. The study also compared various propensity scale methods, finding that the scales by Parker et al. and Levitt performed best. The authors conclude that BepiPred is a reliable tool for predicting linear B-cell epitopes and that further improvements are needed to enhance its predictive power. The method and data sets are publicly available for research purposes.