Benchmark Data

  The benchmark dataset constructed for identifying promoters and their types. It is formed by seven subsets:(1) contains 484 σ²⁴-promoter sequences; (2) contains 134 σ²⁸-promoter sequences; (3) contains 291 σ³²-promoter sequences; (4) contains 163 σ³⁸-promoter sequences; (5) contains 94 σ54-promoter sequences; (6) contains 1,694 σ⁷⁰-promoter sequences; (7) contains 2,860 non-promoter sequences. The length of each sample is 81bp. None of the sequences included here has ≥80% pairwise sequence identity to any other in a same subset. See Eq.1 and the relevant text for further explanation. The benchmark dataset can be downloaded from Supplementary Information S1(pdf,docx).