CONICET | Buscador de Institutos y Recursos Humanos

Many bioinformatics tools have been developed to predict the effect of single nucleotide polymorphisms (SNPs) on gene functionality in an effort to reduce the need for in-vivo assays. The use of these tools in scientific research and even in precision medicine for disease treatment and prevention is very frequent. This situation points out the importance of understanding the semantics of their outputs and the quantification of their possible limitations. However, the large number of tools available and the heterogeneity of their output make their selection, understanding, and comparison a non-trivial task. Most of the works in the literature that compare the tools simplify the problem through the conversion of their output to a binary scale, which reduces the information they provide. In this work, the output consistency of the tools that predict the effect of a mutation in the functionality of a gene is analyzed. A methodology based on ranking indices is proposed. The next step considers the integration of the prediction to improve performance. Six tools frequently used in the literature and available online were selected. Two indices, called K-All and K-Strong, that systematically quantify the differences between outputs are proposed. All possible mutations at the amino acid level on the E2 gene of Chikungunya were evaluated. The indices correctly characterize the tools, yielding similar values for tools with related sources of information. For the inner consistency, K-All varied between 0.39 to 0.63, while K-Strong varied between 0.03 to 0.15. These results reflect the importance of the study of the tools, especially when they are used in genetic tests. Finally, the low levels of consistency should not be necessarily interpreted as a bad result. The diversity between tools is useful to integrate the prediction in a more robust performance.

enviar mensaje