Figure 6
From: PaPI: pseudo amino acid composition to score human protein-coding variants
Data set workflow. Workflow representing the data set selection. Variants from HGMD, 1TGP and ESP were filtered basing on coding, frequency and non-overlapping (unique) variants among the different data sources. In order to evaluate and compare the performances of the variant predictor tools, variants were further filtered for frameshift, stop-disrupting, stop-causing and variants not predictable for the other algorithms.