Abstract
Motivation In silico identification of linear B-cell epitopes represents an important step in the development of diagnostic tests and vaccine candidates, by providing potential high-probability targets for experimental investigation. Current predictive tools were developed under a generalist approach, training models with heterogeneous datasets to develop predictors that can be deployed for a wide variety of pathogens. However, continuous advances in processing power and the increasing amount of epitope data for a broad range of pathogens indicate that training organism or taxon-specific models may become a feasible alternative, with unexplored potential gains in predictive performance.ResultsThis article shows how organism-specific training of epitope prediction models can yield substantial performance gains across several quality metrics when compared to models trained with heterogeneous and hybrid data, and with a variety of widely used predictors from the literature. These results suggest a promising alternative for the development of custom-tailored predictive models with high predictive power, which can be easily implemented and deployed for the investigation of specific pathogens.Availability and implementationThe data underlying this article, as well as the full reproducibility scripts, are available at https://github.com/fcampelo/OrgSpec-paper. The R package that implements the organism-specific pipeline functions is available at https://github.com/fcampelo/epitopes.Supplementary information Supplementary materials are available at Bioinformatics online.
Highlights
In humoral immunity, activated B-lymphocytes (B cells) produce antibodies that bind with specific antigens, and are a key component in vertebrate immune responses (Getzoff et al, 1988; Lodish et al, 2000)
Random Forests are ensemble learning methods that consist of the aggregation of several weaker decision tree (DT) models, with an output based on the combined output of the underlying DTs
Random forests present a good balance between computational cost and performance, and are robust and flexible to work with different data types and scales, which justifies their use in a variety of application domains including several epitope prediction methods (Jespersen et al, 2017; Saravanan and Gautham, 2015)
Summary
In humoral immunity, activated B-lymphocytes (B cells) produce antibodies that bind with specific antigens, and are a key component in vertebrate immune responses (Getzoff et al, 1988; Lodish et al, 2000). The majority of B-cell epitopes are conformational (Van Regenmortel, 1996; Lo et al, 2013), most epitope prediction methods are designed to predict linear epitopes (Alix, 1999; Blythe and Flower, 2005; EL-Manzalawy et al, 2008; Kolaskar and Tongaonkar, 1990; Larsen et al, 2006; Saha and Raghava, 2004, 2006; Singh et al, 2013; Yao et al, 2013) This is mainly due to a relative scarcity of available data on antigen 3D structures, as well as the high computational cost associated with predicting these structures (Yang and Yu, 2009). Discontinuous epitopes can be disrupted by alterations
Sign-in/Register to view highlights & summary 
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call