The strength of the mirscan system is its flexibility and ease of modification. The ability to fluidly modify the features of miRNA hairpins that are examined and the way that they are examined is crucial as our knowledge of miRNA structure and function continues to evolve. The efficiency with which training and evaluation can be performed is also important as miRNAs are examined across the major clades. Scoring matrices developed in one model system may not accurately reflect the properties of miRNAs in another set of species, so the ability to efficiently generate custom scoring systems even when the features being examined remain constant increases this tool's utility. The ability to refine scoring matrices as poor candidates are eliminated also maximizes the power of log-odds scoring in predicting miRNA genes.

The primary shortcoming of the system's current implementation is the inflexibility of the current criteria/feature objects. For instance, the development of independent criteria for examining single-species versus two-species candidate sets allowed me to avoid having to develop truly flexible N-species criteria. Fortunately, the system's flexibility confines the solution for this shortcoming to the development of flexible criteria objects, and makes such solutions more accessible to the comunity by removing the need for re-development of a miRNA-prediction infrastructure to host such improvements.