def __init__(self,filename): self.array=self.toArray(filename) self.numericArray,self.numericColumnIndicesMap,self.numericColumnIndexMap=self.toNumericArray() self.clusters=[] self.enum=self.calcEnumStrs() #prj7: #confidence interval for each species self.cidata=stats.ci(self)
def summarizer(sp_tree, sp_quartet_tree, outfile):
if outfile:
v_out = open(outfile + ".verbose.tsv", "w")
outf = open(outfile + ".tre", "w")
#Data stored on species tree
for i in sp_tree.iternodes():
if i == sp_tree:
continue
#No children means a tip
if len(i.children) == 0:
holder = i.data["qln"]
if outfile:
#convert all floats to strings for printing
temp = list(map(str, holder))
v_out.write(i.label + "\t" + "\t".join(temp) + "\n")
else:
holder = sp_quartet_tree[i.data["q"]]
if outfile:
temp = list(map(str, holder))
v_out.write(
i.get_newick_repr(False) + "\t" + "\t".join(temp) + "\n")
#Make sure something has actually been concordant and met the cutoff
if len(holder) == 0:
mean = 0.0
median = 0.0
min = 0.0
max = 0.0
CIL = 0.0
CIH = 0.0
i.data["concord"] = len(holder)
else:
mean = stats.mean(holder)
median = stats.median(holder)
min = stats.min(holder)
max = stats.max(holder)
i.data["concord"] = len(holder)
#account for the fact you need two for CI's
if len(holder) > 1:
#array and z-value (95% is 1.96)
CIL, CIH = stats.ci(holder, 1.96)
else:
CIL = 0.0
CIH = 0.0
i.data["mean"] = mean
i.data["median"] = median
i.data["min"] = min
i.data["max"] = max
i.data["cih"] = CIH
i.data["cil"] = CIL
array = ["mean", "median", "min", "max", "cil", "cih", "concord"]
for i in array:
if outfile:
outf.write(sp_tree.get_newick_otherlen(i) + ";\n")
else:
print sp_tree.get_newick_otherlen(i) + ";"
'1-sigma': erf(1/sqrt(2)),
'95%': 0.95,
}
z = [int(ki) for ki in logspace(1,3,low_points)]
if mid_points: z += [int(ki) for ki in logspace(3,5,mid_points)]
if high_points: z += [int(ki) for ki in logspace(5,7,high_points)]
z = list(sorted(set(z))) # Make z unique
ni,nj = len(intervals),len(dists)
for pi,interval_name in enumerate(sorted(intervals.keys())):
for pj,dist_name in enumerate(sorted(dists.keys())):
print "processing",dist_name,interval_name
s,rng = intervals[interval_name], dists[dist_name]
#print [min(1000,work//ki+1) for ki in z]
unbiased = [array([ci(rng(ki),s,unbiased=True)
for _ in range(min(maxn,work//ki+1))])
for ki in z]
biased = [array([ci(rng(ki),s,unbiased=False)
for _ in range(min(maxn,work//ki+1))])
for ki in z]
pylab.subplot(ni,nj,pi*nj+pj+1)
#print data
for data in biased, unbiased:
delta = [diff(di,axis=1)[:,0] for di in data]
w = [mean(di) for di in delta]
dw = [std(di) for di in delta]
#print z,w,dw
#print interval_name, dist_name, ni, nj, pi, pj, pj*ni+pi
pylab.semilogx(z,w,'o')
pylab.errorbar(z,w,dw)