Skip to main content

Table 5 Our ISREs predict Ψ more accurately than [9]’s

From: Does conservation account for splicing patterns?

  Our ISREs n = 132 [9] n = 574 Random n = 574
Counts 0.644 ± 0.006 0.601 ± 0.008 0.609 ± 0.008
+NMD 0.810 ± 0.006 0.788 ± 0.005 0.805 ± 0.006
Cons 0.693 ± 0.006 0.608 ± 0.008 0.615 ± 0.007
  1. AUCs for absolute Ψ prediction using k-mer counts, counts plus two nonsense-mediated decay (NMD) features from [12], and conservation-weighted counts for various k-mer sets. The conservation weight of each k-mer instance is the minimum conservation over its k bases, to avoid counting non-functional k-mer instances that partially overlap functional ones. Our ISREs predict Ψ more accurately than [9]’s, which perform no better than the same number of random k-mers