In silico methods for predicting T-cell epitopes: Dr Jekyll or Mr Hyde?

In silico tools offer an attractive alternative strategy to the cumbersome experimental approaches to identify T-cell epitopes. These computational tools have metamorphosed over the years into complex algorithms that attempt to efficiently predict the binding of a plethora of peptides to HLA alleles. In recent years, the scientific community has embraced these techniques to reduce the burden of wet-laboratory experimentation. Although there are some splendid examples of the utility of these methods, there are also evidences where they fall short and remain inconsistent. Hence, are these computational tools 'Dr Jekyll' or 'Mr Hyde' to the researcher, who wishes to utilize them intrepidly? This article reviews the progress and pitfalls of the in silico tools that identify T-cell epitopes.