Classification of seed matches according to their accessibility. All seed matches are divided among bound (B) and unbound (U) sites according to the PAR-CLIP data. Similarly, for a given accessibility threshold Pcutoff and length n of the nucleation region, they are divided into accessible (A) and inaccessible. The fraction of accessible among bound sites (i.e., the conditional probability P(A|B)) is computed as the number of sites that are both bound and accessible (|B ∩ A|) divided by the total number of bound sites (|B|). The fraction of accessible among unbound sites (i.e., the conditional probability P(A|U)) is computed as the number of sites that are both unbound and accessible (|U ∩ A|) divided by the total number of unbound sites (|U|).
