Page 90 - MI-2-2

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

 e i 1   e  i f i   1   f  i g i t i   1      Then, with i microarray data samples, linear Equations
  
Xl  
Yl  
18
 
t i =  l   l  VI-IX can be represented as follows :
i t
  g   l 1 i     
Zl    
 δ 1 l              
   S 1    σ k    ϕ k  
  1
1  
k  
        
2   β
 
 δ   S 2   = σ k    k + ϕ 2 , fork = , ,,K12 … (X)
k  
k  




 XI          
 ζ lI       
     SI     σ     ϕ k   
I I
I  
k   
k   
i
×  + ϕ l   σ   l   * β  ϕ + l  
i =
i  
 ζ YI  l
 − ω        
 1 I       
1  
  1
    t 1    σ l    ϕ l  
l  
       
2   β
 
l  
l  
 − ω ZI   t 2   = σ l    l + ϕ 2 , forl = , ,,L12 … (XI)




 v          
 I       
I  
l   
l   
for i = 1,2,…,L and i=1,2,…,I (VII)   tI     σ II     ϕ l   
i 
 e i    1   e Xm   1   f Ym   g i f  m  

 
i  
 
i f i   
1  
d          
i = 
m  
id 

  g Zm   m   1 i        d  1    σ m      ϕ m   
1  
  1
m  



η    d  2 = σ 2   β + ϕ   2 , form = , ,,M12 … (XII)


 m 1   m     m    m  m   
           
η        
I  
I
I I
m   
m   
 Xm    d     σ     ϕ m   
 κ u 1 
  
i  
i
×  + ϕ  m   σ   m   * β  ϕ + m         
i =
κ  Ym  m      
1  
  1
n  
 − λ   q 1    σ n    ϕ n  





 u 1   q 2   = σ 2   β + ϕ 2 , forn = , ,,N12 … (XIII)
 
   n     n    n  n   
          
 − λ Zm   qI     σ     ϕ I   
I I
 c  m     n      n      n   
for m = 1,2,…,M and i=1,2,…,I (VIII) Equations X-XIII can be rewritten into linear regression
in Equations XIV-XVII.
 e i e   1   f  i g i g i    1   n  
 
 
i f i   
1  
Yn  
Xn  
qi =  n    × S =σ * β k + ∆ k , fork = ,,12 …, K (XIV)
 
k
k
id
  g   m   1 i     
Zn  
 o 1 n  T =σ * β l + ∆ l , forl = ,,12 …, L (XV)
l
l
 
   D =σ * β + ∆ , form = ,,12 …, M (XVI)
 o  m m m m
 Xn 
 ρ 1 n  Q =σ * β n + ∆ n , forn = ,,12 …, N (XVII)
n
n
  
 + ϕ n   σ   n   * β  ϕ + n   where σ , σ, σ , and σ denote the regression vectors for
i
i  
i =
k
 κ Yn  n i protein/gene/miRNA/lncRNA expression data and DNA
n
l
m
 τ − 
 1 n  methylation profiles. β , β , β , and β are the parameter
m
1
k
n
   vectors associated with the protein interactions/gene/
  miRNA/lncRNA regulations, respectively, containing
 τ − Zn 
 u  protein interaction abilities, transcriptional regulatory
 n  abilities, post-transcriptional regulatory abilities, and basal
for n = 1,2,…,N and i=1,2,…,I (IX) levels.
Volume 2 Issue 2 (2025) 82 doi: 10.36922/mi.4620