Algorithm 2 Xprediction: explainable prediction.
1:	Construct prediction models based on the kernel support vector machine (kSVM),
	Random Forest (RF), and Neural Network (NN).
2:	Compute prediction accuracies based on k-fold cross-validation (CV). The average of
	the prediction accuracies of k validation sets was given as: Acc$\left(\widehat{\mathit{y}}\right)$.
3:	Step 2 is iterated N times for randomly constructed k-fold CV datasets.
4:	If $l\le q$, then
5:	If $j\le p$, then
6:	Delete $(l,j)$ elements from regulatory effect matrices: ${\mathit{R}}_{\ell },\ell =1,\dots ,q$
7:	Compuate prediction accuracy of the model without $(l,j)$ elements: Acc$\left({\widehat{\mathit{y}}}^{(l,j)}\right)$.
8:	Step 7 is iterated $N^{(l,j)}$ times for randomly constructed k-fold CV datasets.
9:	Perform t-test between Acc$\left(\widehat{\mathit{y}}\right)$ and Acc$\left({\widehat{\mathit{y}}}^{(l,j)}\right)$ obtained from N and $N^{(l,j)}$ iterations
	and compute p value.
10:	Cruciality of molecular interplays for AI-based prediction results are measured by on
	p value of the t-test.