Page 93 - MI-2-2

Microbes & Immunity Big data and DNN-based DTI model in CHP

regulations, and lncRNA regulations, and epigenetic interactive abilities of proteins, transcriptional regulatory
regulations by DNA methylation, phosphorylation, abilities of TFs and lncRNAs, and post-transcriptional
and ubiquitination of real GENs can be represented in regulatory abilities of miRNAs of the real GWGEN in
Equations XXXIV-XXXVII: Equations XXXIV-XXXVII as the following network
matrix R of real GWGEN (Equation XXXVIII):
W k

Si Si[] []+ β
Si[]= ∑ α kw k w  k  R 0 0 
k
w=1  pp− 
≠
wk  R tg− R lg− R mg 
−
for k = 1,2,…,K and i = 1,2,…,I (XXXIV) R =   R tl− R ll− R ml−   (XXXVIII)
 R tm− R lm− R mm− 
Z l

Y l

X l
ei Pi[] []+
ti[]= ∑ δ lx x x ∑ ζ ly fi Pi[] []− ∑ ω ggi Pi[] []+ vP i[]    
y
l
l
lz
z
l
ll
x=1 y=1 z=1
where submatrix R stands for the matrix associated
p−p
for l = 1,2,…,L and i = 1,2,…,I (XXXV) with the interactive abilities of all proteins as Equation
XXXIX.

X m Y m Z m

di[]= ∑ η mx x ∑ κ my fi[]− ∑ λ mz gi di[] []++ c  α … α … α
ei[]+

y
z
m
m
m
x=1 y=1 z=1  11 1w 1W 
 
 
for m = 1,2,…,M and i = 1,2,…,I (XXXVI)  α 1 k … α kw … α kW  (XXXIX)
  
 

X n Y n Z n α  

qi[]= ∑ oe i[]+ ∑ ρ ny f i[]− ∑ τ nz g iq i[] []++ u n  K 1 … α Kw … α KW 
n
y
z
nx x
n
x=1 y=1 z=1
R , R , and R stand for the matrices associated
t−m
t−l
t−g
with transcriptional regulatory abilities of TFs on genes,
for n = 1,2,…,N and i = 1,2,…,I (XXXVII)
miRNAs, and lncRNAs, respectively, as Equation XL.
where W , X, Y, Z, X , Y , Z , X , Y and Z represent
k
n,
l
l
n
m
n
m
m
l
the interaction number of protein k in PPIs, TF regulations     

on gene l, lncRNA regulations on gene l, miRNA regulations   
on gene l, TF regulations on miRNA (m), lncRNA   
regulations on miRNA (m), miRNA regulations on miRNA   δ 11 … δ 1x … δ 1X  η 11 … η 1x … η 1X  

(m), TF regulations on lncRNA (n), lncRNA regulations    

on lncRNA (n), and miRNA regulations on lncRNA (n) of δ  1 l … δ lx … δ  η  1 … η mx … η mXX


lX

m
real GWGENs obtained by AIC, respectively. α indicates     

kw
 
the estimated protein interactive abilities and δ , η , and  δ δ … δ  η … … η 

…
η
mx
lx
   L 1 Lx LX     M 1 Mx MX  
o denote the estimated transcriptional regulatory abilities
nx


of TFs on gene/miRNA/lncRNA, respectively. ζ , κ ,    
my
ly

and ρ are the estimated transcriptional regulatory abilities    
ny
 

of lncRNAs on gene/miRNA/lncRNA and ω , λ , and τ   o 11 … o x 1 … o 1 X  
nz
mz
lz
represent the estimated post-transcriptional regulatory   (XL)
abilities of miRNAs on gene/miRNA/lncRNA. These o  … o … o 
estimated protein interactive abilities, estimated  n1 nx nX  

transcriptional regulatory abilities, and estimated post-   …  
transcriptional regulatory abilities can be obtained by   o N1 o Nx … o NX  
applying system identification methods in Equations
XVIII-XXI through the corresponding microarray data. R lg− R lm− , and R l−l stand for the matrices associated
After the protein interactive abilities, transcriptional with transcriptional regulatory abilities of LncRNAs on
regulatory abilities, and post-transcriptional regulatory genes, miRNAs, and lncRNAs, respectively, as shown in
abilities of real GWGEN were identified, we integrated all Equation XLI.
Volume 2 Issue 2 (2025) 85 doi: 10.36922/mi.4620