Comparison of predictions across datasets (red and pink symbols) and within datasets is borderline significant.
← all excerpts
Impact of between-tissue differences on pan-cancer predictions of drug sensitivity.
2
—
—
The sentences
When comparing within- and cross-MEKi predictions, we found marginally significant better performance for predictions of the same MEKi (ρ mean: 0.59; range: 0.47–0.65, circles in Fig 2D ) than those for different MEKi (ρ mean: 0.55; range: 0.46–0.64, triangles) (U test, p = 0.10).