def all_corr(vec):
ln=len(vec[0])
print ln, ln
print len(vec), len(vec[0])
vec=np.array(vec)
norm=[]
for ii in xrange(ln):
norm.append(ss.zscore(vec[:,ii]))
norm=analyse.data2arr(norm)
print 'spearman ...'
#cor_mat=ss.spearmanr(norm)[0]
cor_mat=[]
for ii in range(len(vec)):
cor_mat.append([])
for jj in xrange(len(vec)):
tmp=ss.pearsonr(vec[ii],vec[jj])
cor_mat[-1].append(tmp[0])
#print len(vec[:,ii])
cor_mat=np.array(cor_mat)
####
dist=1-cor_mat
distance=[]
for i in xrange(len(dist)):
distance=distance+list(dist[i][i+1:])
print len(distance)
distance=np.array(distance)
return norm.T,cor_mat,distance
def CPG_RPKM(RPKM, CPG, lim):
CGI=[]
cpg=read.read_dat(CPG,'\t')
for i in cpg:
CGI.append(i[0])
genes=RPKM[0][1:]
libs=[]
mat=[]
for i in RPKM[1:]:
mat.append(i[1:])
libs.append(i[0])
mat=analyse.data2arr(mat)
genes=np.array(genes)
allave=[]
cgiave=[]
print len(mat)
for i in range(len(mat)):
gntmp=genes[mat[i,:]>lim]
tmp=mat[i,:][mat[i,:]>lim]
allave.append(np.mean(tmp))
cgirpkm=[]
for j in xrange(len(gntmp)):
if gntmp[j] in CGI:
cgirpkm.append(tmp[j])
cgiave.append(np.mean(cgirpkm))
print len(gntmp), len(cgirpkm)
allave=np.array(allave)
cgiave=np.array(cgiave)
return allave,cgiave, libs
def CPG_RPKM(RPKM, CPG, lim):
CGI = []
cpg = read.read_dat(CPG, '\t')
for i in cpg:
CGI.append(i[0])
genes = RPKM[0][1:]
libs = []
mat = []
for i in RPKM[1:]:
mat.append(i[1:])
libs.append(i[0])
mat = analyse.data2arr(mat)
genes = np.array(genes)
allave = []
cgiave = []
print len(mat)
for i in range(len(mat)):
gntmp = genes[mat[i, :] > lim]
tmp = mat[i, :][mat[i, :] > lim]
allave.append(np.mean(tmp))
cgirpkm = []
for j in xrange(len(gntmp)):
if gntmp[j] in CGI:
cgirpkm.append(tmp[j])
cgiave.append(np.mean(cgirpkm))
print len(gntmp), len(cgirpkm)
allave = np.array(allave)
cgiave = np.array(cgiave)
return allave, cgiave, libs
def all_corr(vec):
ln = len(vec[0])
print ln, ln
print len(vec), len(vec[0])
vec = np.array(vec)
norm = []
for ii in xrange(ln):
norm.append(ss.zscore(vec[:, ii]))
norm = analyse.data2arr(norm)
print 'spearman ...'
#cor_mat=ss.spearmanr(norm)[0]
cor_mat = []
for ii in range(len(vec)):
cor_mat.append([])
for jj in xrange(len(vec)):
tmp = ss.pearsonr(vec[ii], vec[jj])
cor_mat[-1].append(tmp[0])
#print len(vec[:,ii])
cor_mat = np.array(cor_mat)
####
dist = 1 - cor_mat
distance = []
for i in xrange(len(dist)):
distance = distance + list(dist[i][i + 1:])
print len(distance)
distance = np.array(distance)
return norm.T, cor_mat, distance
def data2matrix(T,NA): x=T[0][1:] y=[] mat=[] for i in xrange(len(T)): for j in xrange(len(T[0])): if T[i][j]=='NA' or T[i][j]=='N/A': T[i][j]=NA for i in T[1:]: y.append(i[0]) mat.append(map(float,i[1:])) #print mat[0] mat=analyse.data2arr(mat) return x,y,mat
