Figure 2.

Methodology of inclusion of acquired knowledge (InAKnow). According to the ribosome scanning model, the sequences extracted from the downstream region of the TIS do not have classification. Thus, at Level 1 of this proposed methodology, the sequences of the upstream region (positive and negative) are trained by the classifier, and at Level 2 the model created by this training is applied to the sequences of the downstream region (without classification). Subsequently, the sequences of the downstream region (now classified), together with those of the upstream region are again inserted into new training, thereby generating a new model (Level 3).

Silva et al. BMC Genomics 2011 12(Suppl 4):S9   doi:10.1186/1471-2164-12-S4-S9