def updateEDMSTI(d, seq, tree, num, clusteringType):
newD = {}
if clusteringType == 'NJ':
maxSimilarity = NJUtils.getMaxSimilarityForNJ(d, seq)
elif clusteringType == 'UPGMA':
maxSimilarity = UPGMAUtils.getMaxSimilarityForUPGMA(d)
elif clusteringType == 'WPGMA':
maxSimilarity = WPGMAUtils.getMaxSimilarityForWPGMA(d)
f = maxSimilarity['key'][0]
iF = seq.index(f)
numF = num[iF]
g = maxSimilarity['key'][1]
if f == g:
iG = iF + 1 + seq[iF + 1:].index(g)
else:
iG = seq.index(g)
numG = num[iG]
u = '(' + f + ':' + str(d[maxSimilarity['key']]) + ':' + g + ')'
seq.remove(f)
seq.remove(g)
if clusteringType == 'NJ':
delta = NJUtils.getDeltaDistForNJ(d, f, g, u, seq)
elif clusteringType == 'UPGMA':
delta = UPGMAUtils.getDeltaDistForUPGMA(d, f, g, u)
elif clusteringType == 'WPGMA':
delta = WPGMAUtils.getDeltaDistForWPGMA(d, f, g, u)
tree.append({'1': u, '2': f, 'dist': delta[u, f]})
tree.append({'1': u, '2': g, 'dist': delta[u, g]})
for i in range(len(seq)):
for j in range(len(seq)):
if i != j:
newD[seq[i], seq[j]] = d[seq[i], seq[j]]
if clusteringType == 'NJ':
newD = NJUtils.getDistUForNJ(d, newD, f, g, u, seq)
elif clusteringType == 'UPGMA':
newD = UPGMAUtils.getDistUForUPGMA(d, newD, f, g, u, seq, numF, numG)
elif clusteringType == 'WPGMA':
newD = WPGMAUtils.getDistUForWPGMA(d, newD, f, g, u, seq)
seq.append(u)
num.append(numF + numG + 1)
return newD, seq, num, tree, iF, iG
def nj(distMat, texaName = None, precision = 5, dspTree = True, saveFlag = True):
"""
unrooted tree, last three clusters
"""
#check input distance matrix format
distMat = np.array(distMat, dtype = float);
util.checkInput(distMat);
# Initialization
#njTree = JTree;
distMatInput = distMat[:] #deep copy for estimate error calculation
numTexa = distMat.shape[0];
if texaName:
newick = texaName[:];
else:
newick = [ch for ch in string.ascii_uppercase[0:numTexa]];
texaName = newick[:];
nodeName = newick[:]; # cluster represented by each node in nodeID
nodePT = range(0, numTexa); # node pointer
edgeLengths = [];
# combine clusters till the number of clusters (nc) is two
for nc in range(numTexa, 2, -1):
colSum = np.sum(distMat, axis = 0); # sums of columns
Q = (nc - 2) * distMat - np.tile(np.matrix(colSum).T, (1, nc)) \
- np.tile(colSum, (nc, 1)) \
+ np.diag(np.repeat(np.inf, nc)); #Studier & Keppler
# find the indicies corresponding to the minimum value of Q
# consider only upper triangle to make sure i > j format of indices
# print "Q", (Q == Q.T).all()
[i,j] = util.findMinInd(Q, nodePT, numTexa, precision);
nodeJoin = [nodePT[i], nodePT[j]];
# Update node pointer, node name, node ID
nodeName.append("(" + nodeName[nodePT[i]]
+ "," + nodeName[nodePT[j]] + ")");
newNodeInd = range(0,i) + range(i+1,j) + range(j+1,nc);
nodePT = [nodePT[dummy] for dummy in newNodeInd];
nodePT.append(len(nodeName) - 1);
# compute distance from i and j to the new node (ij)
branchLen = (np.dot(colSum[[i,j]], np.array([[1,-1],[-1,1]]))/(2*(nc-2))
+ distMat[i,j]/2);
edgeLengths.extend(branchLen);
# update newick string
newickStr = util.newickStringFormat(newick, nodeJoin,
branchLen, precision);
newick.append(newickStr);
# compute distance between the new node (ij) and remaining nodes
# excluding i and j and construct new distance matrix.
dist = (np.sum(distMat[np.ix_([i,j], newNodeInd)], axis = 0)
- distMat[i,j]) / 2;
dist = np.array([dist]);
# if np.any(dist < 0):
# dist = np.fmax(dist,0); #no negative distance
distMatTop = np.concatenate((distMat[np.ix_(newNodeInd, newNodeInd)],
dist.T), axis = 1);
distMatBottom = np.append(dist, 0);
distMat = np.vstack((distMatTop,[distMatBottom])); # update distance matrix
nodeName.append("(" + nodeName[nodePT[0]]
+ "," + nodeName[nodePT[1]] + ")");
newickStr = util.newickStringFormat(newick, nodePT,
distMat[[0,1],[1,0]]/2, precision) + ";";
newick.append(newickStr);
edgeLengths.append(distMat[0,1]);
print "Tree Building\n", nodeName
print "\nNewick Tree\n", newick[-1]
if dspTree:
import rpy2.robjects.packages as rp
tree = Phylo.read(StringIO(newick[-1]), "newick");
Phylo.draw(tree);
ape = rp.importr('ape')
t = ape.read_tree(text = newick[-1])
ape.plot_phylo(t, type = 'unrooted')
util.estimateError(newick[-1], distMatInput, texaName);
print "Total length of the tree L = ", sum(edgeLengths)
if saveFlag:
with open("result.nw", 'w') as f:
f.write(newick[-1]);
return newick[-1]
Phylo.draw(tree);
ape = rp.importr('ape')
t = ape.read_tree(text = newick[-1])
ape.plot_phylo(t, type = 'unrooted')
util.estimateError(newick[-1], distMatInput, texaName);
print "Total length of the tree L = ", sum(edgeLengths)
if saveFlag:
with open("result.nw", 'w') as f:
f.write(newick[-1]);
return newick[-1]
testData = util.importDistMat('testData/TRex/distmat_6.dat');
names = testData[0];
distMat = testData[1];
test = nj(distMat, names);
###
#import test_data
#dataNum = 1;
#distMat = test_data.distMat[dataNum]
#if dataNum in test_data.names:
# names = test_data.names[dataNum];
# test = nj(distMat, names);
#else:
# test = nj(distMat)
#distMat = util.randomDistMatGen(25);
#test = nj(distMat)
