Figure 2 | Scientific Reports

Figure 2

From: Understanding the hidden relations between pro- and anti-inflammatory cytokine genes in bovine oviduct epithelium using a multilayer response surface method

Figure 2

A representative structure of the MLRSM scenario 2 in predicting IL1B mRNA expression. (a) In the first calibrating process, a high-nonlinear polynomial function was employed to calibrate the hidden database (the predicted gene pairs) (y) by using the input dataset (x), i.e., TNFA, TLR4, IL10, and IL4, regardless to the experimental conditions. (b) The hidden database was shown as the predicted gene pairs, i.e., the co-expression of (TNFA + TLR4), (IL10 + TLR4), (TNFA + IL10), (TLR4 + IL4), (TNFA + IL4), or (IL10 + IL4). Next, in the second calibrating process, the handling predicted data (the hidden layer, the predicted gene pairs) from the first calibrating process were employed for the regression analysis of the target (output) gene, G, (i.e., IL1B). (c) Due to the inclusion of the hidden layer to the first modeling approach; the model was able to consider the correlation between the input genes (x). Therefore, we utilized the data of the hidden layer (the predicted gene pairs) to construct the relations between gene pairs and output gene, i.e., IL1B, as seen in Fig. 5a–e).

Back to article page