“…First, is the need to identify the two important parameters in GRN construction that are affected by noise: (1) the unaffected genes and (2) the wild-type strain values, which are more difficult to identify when a larger number of genes are involved. Second, though past research has been conducted in reconstructing GRN, only a few researchers applied their methods to real experimental GRN datasets, as an addition to synthetic data: [30], [41], [40], [43], [5], [44], [45], [46], [47], [21] and [6]. Third, most past research have focused on GRN prediction, with only minor attention given to determining the directionality of the genes.…”