Skip to main content

Table 3 Pair-wise overlaps of correct predicted promoters between two PPPs at medium (200 bp) resolution for the CpG-rich promoters.

From: MetaProm: a neural network based meta-predictor for alternative human promoter prediction

  PPP (correct predictions)
   DragonGSF (7,205) DragonPF (18,959) FirstEF (15,015) FProm (20,884) McProm (7,891)
PPP (correct predictions) DragonPF (18,959) 49% (6,380)     
  FirstEF (15,015) 44% (4,890) 75% (12,795)    
  FProm (20,884) 45% (6,310) 81% (16,084) 70% (12,632)   
  McProm (7,891) 38% (2,870) 52% (6,918) 49% (5,636) 49% (7,051)  
  PSPA (11,164) 47% (4,313) 67% (10,091) 67% (8,750) 61% (9,806) 48% (4,534)
  1. The number in parenthesis is the count of correctly predicted promoters; the percentage number is calculated by the count of the overlap divided by the mean of both PPPs. The overlaps at high and low resolutions are shown in Additional file 3.