abhik1368

create table tmp_data as 
SELECT DISTINCT molecule_dictionary.chembl_id,compound_structures.canonical_smiles,ASSAYS.TID,target_dictionary.chembl_id AS target_chembl_id,
TARGET_DICTIONARY.ORGANISM,ACTIVITIES.*,assays.chembl_id AS assay_chembl_id 
from molecule_dictionary,compound_structures,assays,activities,target_dictionary 
where compound_structures.molregno = MOLECULE_DICTIONARY.MOLREGNO 
and MOLECULE_DICTIONARY.MOLREGNO = ACTIVITIES.MOLREGNO 
and ASSAYS.TID = TARGET_DICTIONARY.TID AND ASSAYS.ASSAY_ID = ACTIVITIES.ASSAY_ID 
and ACTIVITIES.STANDARD_UNITS = 'nM' AND ACTIVITIES.STANDARD_VALUE < 50 
and TARGET_DICTIONARY.ORGANISM = 'Homo sapiens' 

abhik1368

install.packages('RPostgreSQL', type='source')

#initiate connection
library(RPostgreSQL)
library(BMS)
library(ggplot2)

drv <- dbDriver("PostgreSQL")
con <- dbConnect(drv, dbname="chembl_18")

abhik1368

# Create function
create or replace function inchi_mol(x text)
returns text
AS $$
from rdkit import Chem
from rdkit.Chem import AllChem
file = open("/home/chembl/logerror2.txt","a")
try :
    m=Chem.MolFromInchi(x)
    if m is not None :

abhik1368

# x = a vector for scores
# y = a vector of labels

function (x, y, decreasing = TRUE, top = 1) 
{
    if (length(x) != length(y)) {
        stop(paste("Length of scores does not match with labels."))
    }
    N <- length(y)
    n <- sum(y == 1)

abhik1368

# x = a vector of scores 
# y = a vector of labels

function (x, y, decreasing = TRUE, alpha = 20) 
{
    if (length(x) != length(y)) {
        stop(paste("The length of scores should be equal to number of labels."))
    }
    N <- length(y)
    n <- length(which(y == 1))

abhik1368

# x = a vector of scores
# y = a vector of labels

function (x, y, decreasing = TRUE, alpha = 20) 
{
    if (length(x) != length(y)) {
        stop(paste("The length of scores should be equal to number of labels."))
    }
    N <- length(y)
    n <- length(which(y == 1))

abhik1368

# Parameter r: restart probability
r<-0.8
# convergence cutoff 
conv_cut<-1e-10

RWR <- function(M, p_0, r,conv_cut, prop=FALSE) {
    
    # use Network propagation when prop=TRUE
    if(prop) {
        w<-colSums(M)

abhik1368

# Algorithm to perform MCL Clustering in R 
# Add the identity matrix to the matrix which indicates self loops 
add.selfloops <- function (M) {
  LM<-M+diag(dim(M)[1])
  return (LM);
}

# Inflation step of MCL
inflate <- function (M,inf) {
  M <- M^(inf)

abhik1368

nbi < - function (A){
    # A is the n x m adjacencny matrix here
    n <- nrow(A)
    m <- ncol(A)
    # You need to calculate the degree of columns to use it as node weight
    Ky <- diag(1/colSums(A))
    Ky[is.infinite(Ky) | is.na(Ky)] <- 0
    kx <- rowSums(A)
    Nx <- 1/(matrix(kx, nrow=n, ncol=n, byrow=TRUE))

abhik1368

data(Enzyme)

# Adjacency Matrix

A = enzyme_ADJ
# Compute inner product
m <- t(A) %*% A
m[m>0] <- 1

# Chemical similarity Matrix