Fig. 1
From: Benchmarking splice variant prediction algorithms using massively parallel splicing assays
Variants used for splice effect predictor benchmarking. A Validation sets can be drawn from pathogenic clinical variants and, conversely, common polymorphisms in frequently screened disease genes (top panel), broadly targeted massively parallel splice assays (MPSAs) interrogating a few variants across many exons (middle panel), and saturation MPSAs in which all possible variants are created for a few target exons (bottom panel). B Variant classes defined by exon/intron region and proximity to splice sites (upper), with the percent coverage of the possible SNVs within each variant class (denoted by color) for each dataset in the benchmark set (for BRCA1, missense, and stop-gain variants were excluded and not counted in the denominator)