Skip to main content
Fig. 2 | BMC Research Notes

Fig. 2

From: Social networks help to infer causality in the tumor microenvironment

Fig. 2

Algorithm description in three steps. In the first step, a co-expression network is constructed based on the calculation of direct and indirect correlation. Direct correlation refers to the second-order partial correlation, whereas indirect correlation is calculated as the difference between zeroth-order partial correlation and the direct correlation. In the second step, a bidirectional network of influence is constructed. Forward and reverse influences are derived from the calculation of the investments for each couple of genes. The directionality assignment of the investments is based on the slope ratio (SR) criterion. In the third step, causality is inferred by evaluating every couple of forward and reverse influences. The directionality of the predictions is the opposite of the highest influence value and can be interpreted as the directionality of the main flux of regulatory effects

Back to article page