def _gapStatus(self): ### Predicts ancestral gap status
'''Predicts ancestral gap status.'''
try:
### Setup ###
tree = self.obj['Tree']
gaspnode = self.gaspnode
progress = 0.01 * 2 * len(tree.node[tree.stat['SeqNum']:]) + len(gaspnode[tree.node[-1]].sequence)
p = 0
### From tips to root ###
self.log.printLog('\r#GASP','Calculating Gap Status (Down) ... 0.0%',log=False,newline=False)
for node in tree.node[tree.stat['SeqNum']:]:
an = gaspnode[node]
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
n = [gaspnode[desc[0]],gaspnode[desc[1]]]
for r in range(len(an.sequence)):
an.ancgap[r] = (n[0].ancgap[r] + n[1].ancgap[r])/2
an.nodegap = 1
p += 100.0
self.log.printLog('\r#GASP','Calculating Gap Status (Down) ... %.1f%%' % (p/progress),log=False,newline=False)
### Fix Root ###
an = gaspnode[tree.node[-1]]
for r in range(len(an.sequence)):
if an.ancgap[r] <= 0.5:
an.sequence = rje.strSub(an.sequence,r,r,'X')
an.ancgap[r] = 0.0
else:
an.sequence = rje.strSub(an.sequence,r,r,'-')
an.ancgap[r] = 1.0
p += 100.0
self.log.printLog('\r#GASP','Calculating Gap Status (Root) ... %.1f%%' % (p/progress),log=False,newline=False)
### Back up tree using all 3 branches ###
uptree = tree.node[tree.stat['SeqNum']:-1]
uptree.reverse()
for node in uptree:
an = gaspnode[node]
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
n = [gaspnode[desc[0]],gaspnode[desc[1]],gaspnode[anc]]
for r in range(len(an.sequence)):
an.ancgap[r] = (n[0].ancgap[r] + n[1].ancgap[r] + n[2].ancgap[r])/3
if an.ancgap[r] <= 0.5:
an.sequence = rje.strSub(an.sequence,r,r,'X')
an.ancgap[r] = 0.0
else:
an.sequence = rje.strSub(an.sequence,r,r,'-')
an.ancgap[r] = 1.0
p += 100.0
self.log.printLog('\r#GASP','Calculating Gap Status ( Up ) ... %.1f%%' % (p/progress),log=False,newline=False)
self.log.printLog('\r#GASP','Calculation of Gap Status complete. ')
except:
self.log.errorLog('Fatal Error during _gapStatus().')
raise
def addPhos(self,seq,pos,type): ### Adds relevant marker to given sequence
'''Adds relevant marker to given sequence.'''
hierarchy = ['*','?','H','M','L','-']
markers = {'Hom':'?','ID':'*','Self':'*'}
p = markers[type]
x = seq.info['Sequence'][pos]
if x not in hierarchy or hierarchy.index(x) > hierarchy.index(p):
seq.info['Sequence'] = rje.strSub(seq.info['Sequence'],pos,pos,p)
def addPhos(self, seq, pos,
type): ### Adds relevant marker to given sequence
'''Adds relevant marker to given sequence.'''
hierarchy = ['*', '?', 'H', 'M', 'L', '-']
markers = {'Hom': '?', 'ID': '*', 'Self': '*'}
p = markers[type]
x = seq.info['Sequence'][pos]
if x not in hierarchy or hierarchy.index(x) > hierarchy.index(p):
seq.info['Sequence'] = rje.strSub(seq.info['Sequence'], pos, pos,
p)
def _gaspProbs(self,
aalist,
useanc=False,
dir='down',
aaprobs=True,
aasub=False,
aafix=False,
gpass=0): ### Work through tree calculating AA probs
'''
Work through tree calculating AA probs.
>> aalist:list = alphabet
>> useanc:boolean [False] = whether to consider predicted ancestor
>> dir:str ['down'] = direction to move through tree ('up','down','rootonly')
>> aaprobs:boolean [True] = whether to calculate AA probabilities
>> aasub:boolean [False] = whether to change sequence to most likely AA
>> aafix:boolean [False] = whether to fix most likely AA as probability 1.0
>> gpass:int [0] = Extra pass (for log output only)
'''
try:
### Setup ###
tree = self.obj['Tree']
seqlen = tree.node[0].obj['Sequence'].seqLen()
forloop = tree.node[tree.stat['SeqNum']:]
if dir == 'up':
forloop.reverse()
forloop = forloop[1:]
elif dir == 'root':
forloop = [forloop[-1]]
logtxt = '(%s)' % dir
if aasub:
logtxt = 'substitutions (%s)' % dir
elif aaprobs:
logtxt = 'probabilities (%s)' % dir
if gpass > 0:
logtxt = '[Extra Pass %d] %s' % (gpass, logtxt)
rstseq = [''
] * seqlen # RST sequences for each column of alignment
for node in tree.node[:tree.stat['SeqNum']]:
for r in range(seqlen):
rstseq[r] += node.obj['Sequence'].info['Sequence'][r]
### <b> ### Each gaspnode in turn
for node in forloop:
self.log.printLog('\r#GASP',
'Calculating GASP %s: %.1f%%' %
(logtxt, 100.0 * float(forloop.index(node)) /
len(forloop)),
log=False,
newline=False)
an = self.gaspnode[node]
if node == tree.node[-1]: # Root
desc_seq = tree.node
else:
clades = tree.branchClades(node.ancBranch())
desc_seq = clades[1]
## <i> ## AA probabilities
if aaprobs:
# Establish relevant nodes and PAM distances
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
treenodes = desc
n = [self.gaspnode[desc[0]],
self.gaspnode[desc[1]]] # Gasp nodes
p = {}
if anc != None:
treenodes.append(anc) # Tree Nodes
n.append(self.gaspnode[anc])
for tn in treenodes:
p[tn] = self.stat['FixPam'] # pam distances
if self.stat['FixPam'] == 0:
p[tn] = int(node.link(tn).stat['Length'] * 100) + 1
# Calculate Probabilities for each residue
bigproblem = False
for r in range(seqlen):
if (an.ancgap[r]
== 0) and (an.ancfix[r]
== False): # No gap & not fixed
if n[0].ancfix[r] and n[1].ancfix[r] and (
n[0].sequence[r] == n[1].sequence[r]
): # Fixed in descendants
a = n[0].sequence[r]
an.sequence = rje.strSub(an.sequence, r, r, a)
an.ancaap[r][a] = 1.0
an.ancfix[r] = True
else: # Variation
for a in aalist:
aap = [0.0] * 3
#print a, aap, 'PAM', p[desc[0]], p[desc[1]]
for d in aalist:
#print a, '->', d, 'PAM', p[desc[0]], tree.obj['PAM'].matrix[p[desc[0]]].pamp[a+d], '*' ,n[0].ancaap[r][d]
try:
aap[0] += tree.obj['PAM'].matrix[p[
desc[0]]].pamp[
a + d] * n[0].ancaap[r][d]
except:
self.errorLog(
'Big Problem with GASP Prob calculation',
quitchoice=True)
raise
if bigproblem:
self.errorLog(
'Big Problem with GASP Prob calculation'
)
else:
bigproblem = True
#print a, '->', d, 'PAM', p[desc[1]], tree.obj['PAM'].matrix[p[desc[1]]].pamp[a+d], '*' ,n[1].ancaap[r][d]
try:
aap[1] += tree.obj['PAM'].matrix[p[
desc[1]]].pamp[
a + d] * n[1].ancaap[r][d]
except:
self.errorLog(
'Big Problem with GASP Prob calculation',
quitchoice=True)
raise
if bigproblem:
self.errorLog(
'Big Problem with GASP Prob calculation'
)
else:
bigproblem = True
try:
if useanc and anc != None:
aap[2] += tree.obj[
'PAM'].matrix[p[anc]].pamp[
d +
a] * n[2].ancaap[r][d]
except:
self.errorLog(
'Big Problem with GASP Prob calculation',
quitchoice=True)
raise
if bigproblem:
self.errorLog(
'Big Problem with GASP Prob calculation'
)
else:
bigproblem = True
if useanc and anc != None:
an.ancaap[r][a] = (aap[0] + aap[1] +
aap[2]) / 3
else:
an.ancaap[r][a] = (aap[0] + aap[1]) / 2
# Adjust Probabilities to total 1, using RareCut and DescOnly
if dir != 'up':
desc_seq = tree.node
an.adjustAncAAP(self.stat['RareCut'],
desconly=self.opt['DescOnly'],
desc=desc_seq)
if self.opt['RST']:
an.makeRST(rstseq)
if aasub:
for r in range(seqlen):
if (an.sequence[r] != '-') & (an.ancfix[r] == False):
maxp = 0.0
for a in aalist:
if an.ancaap[r][a] > maxp:
an.sequence = rje.strSub(
an.sequence, r, r, a)
maxp = an.ancaap[r][a]
elif (an.ancaap[r][a]
== maxp) & (self.aafreq[a] >
self.aafreq[an.sequence[r]]):
an.sequence = rje.strSub(
an.sequence, r, r, a)
if maxp == 0: # No AA => X
an.sequence = rje.strSub(
an.sequence, r, r, 'X')
if aafix:
for a in aalist:
if a == an.sequence[r]:
an.ancaap[r][a] = 1.0
else:
an.ancaap[r][a] = 0.0
### Finish ###
self.log.printLog('\r#GASP',
'Calculated GASP %s: 100.0%%.' % logtxt)
except:
self.log.errorLog('Fatal Error during _gaspProbs().')
raise
def _gapStatus(self): ### Predicts ancestral gap status
"""Predicts ancestral gap status."""
try:
### Setup ###
tree = self.obj["Tree"]
gaspnode = self.gaspnode
progress = 0.01 * 2 * len(tree.node[tree.stat["SeqNum"] :]) + len(gaspnode[tree.node[-1]].sequence)
p = 0
### From tips to root ###
self.log.printLog("\r#GASP", "Calculating Gap Status (Down) ... 0.0%", log=False, newline=False)
for node in tree.node[tree.stat["SeqNum"] :]:
an = gaspnode[node]
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
n = [gaspnode[desc[0]], gaspnode[desc[1]]]
for r in range(len(an.sequence)):
an.ancgap[r] = (n[0].ancgap[r] + n[1].ancgap[r]) / 2
an.nodegap = 1
p += 100.0
self.log.printLog(
"\r#GASP", "Calculating Gap Status (Down) ... %.1f%%" % (p / progress), log=False, newline=False
)
### Fix Root ###
an = gaspnode[tree.node[-1]]
for r in range(len(an.sequence)):
if an.ancgap[r] <= 0.5:
an.sequence = rje.strSub(an.sequence, r, r, "X")
an.ancgap[r] = 0.0
else:
an.sequence = rje.strSub(an.sequence, r, r, "-")
an.ancgap[r] = 1.0
p += 100.0
self.log.printLog(
"\r#GASP", "Calculating Gap Status (Root) ... %.1f%%" % (p / progress), log=False, newline=False
)
### Back up tree using all 3 branches ###
uptree = tree.node[tree.stat["SeqNum"] : -1]
uptree.reverse()
for node in uptree:
an = gaspnode[node]
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
n = [gaspnode[desc[0]], gaspnode[desc[1]], gaspnode[anc]]
for r in range(len(an.sequence)):
an.ancgap[r] = (n[0].ancgap[r] + n[1].ancgap[r] + n[2].ancgap[r]) / 3
if an.ancgap[r] <= 0.5:
an.sequence = rje.strSub(an.sequence, r, r, "X")
an.ancgap[r] = 0.0
else:
an.sequence = rje.strSub(an.sequence, r, r, "-")
an.ancgap[r] = 1.0
p += 100.0
self.log.printLog(
"\r#GASP", "Calculating Gap Status ( Up ) ... %.1f%%" % (p / progress), log=False, newline=False
)
self.log.printLog("\r#GASP", "Calculation of Gap Status complete. ")
except:
self.log.errorLog("Fatal Error during _gapStatus().")
raise
def _gaspProbs(
self, aalist, useanc=False, dir="down", aaprobs=True, aasub=False, aafix=False, gpass=0
): ### Work through tree calculating AA probs
"""
Work through tree calculating AA probs.
>> aalist:list = alphabet
>> useanc:boolean [False] = whether to consider predicted ancestor
>> dir:str ['down'] = direction to move through tree ('up','down','rootonly')
>> aaprobs:boolean [True] = whether to calculate AA probabilities
>> aasub:boolean [False] = whether to change sequence to most likely AA
>> aafix:boolean [False] = whether to fix most likely AA as probability 1.0
>> gpass:int [0] = Extra pass (for log output only)
"""
try:
### Setup ###
tree = self.obj["Tree"]
seqlen = tree.node[0].obj["Sequence"].seqLen()
forloop = tree.node[tree.stat["SeqNum"] :]
if dir == "up":
forloop.reverse()
forloop = forloop[1:]
elif dir == "root":
forloop = [forloop[-1]]
logtxt = "(%s)" % dir
if aasub:
logtxt = "substitutions (%s)" % dir
elif aaprobs:
logtxt = "probabilities (%s)" % dir
if gpass > 0:
logtxt = "[Extra Pass %d] %s" % (gpass, logtxt)
rstseq = [""] * seqlen # RST sequences for each column of alignment
for node in tree.node[: tree.stat["SeqNum"]]:
for r in range(seqlen):
rstseq[r] += node.obj["Sequence"].info["Sequence"][r]
### <b> ### Each gaspnode in turn
for node in forloop:
self.log.printLog(
"\r#GASP",
"Calculating GASP %s: %.1f%%" % (logtxt, 100.0 * float(forloop.index(node)) / len(forloop)),
log=False,
newline=False,
)
an = self.gaspnode[node]
if node == tree.node[-1]: # Root
desc_seq = tree.node
else:
clades = tree.branchClades(node.ancBranch())
desc_seq = clades[1]
## <i> ## AA probabilities
if aaprobs:
# Establish relevant nodes and PAM distances
anc = node.ancNode()
desc = node.neighbours(ignore=[anc])
treenodes = desc
n = [self.gaspnode[desc[0]], self.gaspnode[desc[1]]] # Gasp nodes
p = {}
if anc != None:
treenodes.append(anc) # Tree Nodes
n.append(self.gaspnode[anc])
for tn in treenodes:
p[tn] = self.stat["FixPam"] # pam distances
if self.stat["FixPam"] == 0:
p[tn] = int(node.link(tn).stat["Length"] * 100) + 1
# Calculate Probabilities for each residue
bigproblem = False
for r in range(seqlen):
if (an.ancgap[r] == 0) and (an.ancfix[r] == False): # No gap & not fixed
if (
n[0].ancfix[r] and n[1].ancfix[r] and (n[0].sequence[r] == n[1].sequence[r])
): # Fixed in descendants
a = n[0].sequence[r]
an.sequence = rje.strSub(an.sequence, r, r, a)
an.ancaap[r][a] = 1.0
an.ancfix[r] = True
else: # Variation
for a in aalist:
aap = [0.0] * 3
# print a, aap, 'PAM', p[desc[0]], p[desc[1]]
for d in aalist:
# print a, '->', d, 'PAM', p[desc[0]], tree.obj['PAM'].matrix[p[desc[0]]].pamp[a+d], '*' ,n[0].ancaap[r][d]
try:
aap[0] += tree.obj["PAM"].matrix[p[desc[0]]].pamp[a + d] * n[0].ancaap[r][d]
except:
self.errorLog("Big Problem with GASP Prob calculation", quitchoice=True)
raise
if bigproblem:
self.errorLog("Big Problem with GASP Prob calculation")
else:
bigproblem = True
# print a, '->', d, 'PAM', p[desc[1]], tree.obj['PAM'].matrix[p[desc[1]]].pamp[a+d], '*' ,n[1].ancaap[r][d]
try:
aap[1] += tree.obj["PAM"].matrix[p[desc[1]]].pamp[a + d] * n[1].ancaap[r][d]
except:
self.errorLog("Big Problem with GASP Prob calculation", quitchoice=True)
raise
if bigproblem:
self.errorLog("Big Problem with GASP Prob calculation")
else:
bigproblem = True
try:
if useanc and anc != None:
aap[2] += tree.obj["PAM"].matrix[p[anc]].pamp[d + a] * n[2].ancaap[r][d]
except:
self.errorLog("Big Problem with GASP Prob calculation", quitchoice=True)
raise
if bigproblem:
self.errorLog("Big Problem with GASP Prob calculation")
else:
bigproblem = True
if useanc and anc != None:
an.ancaap[r][a] = (aap[0] + aap[1] + aap[2]) / 3
else:
an.ancaap[r][a] = (aap[0] + aap[1]) / 2
# Adjust Probabilities to total 1, using RareCut and DescOnly
if dir != "up":
desc_seq = tree.node
an.adjustAncAAP(self.stat["RareCut"], desconly=self.opt["DescOnly"], desc=desc_seq)
if self.opt["RST"]:
an.makeRST(rstseq)
if aasub:
for r in range(seqlen):
if (an.sequence[r] != "-") & (an.ancfix[r] == False):
maxp = 0.0
for a in aalist:
if an.ancaap[r][a] > maxp:
an.sequence = rje.strSub(an.sequence, r, r, a)
maxp = an.ancaap[r][a]
elif (an.ancaap[r][a] == maxp) & (self.aafreq[a] > self.aafreq[an.sequence[r]]):
an.sequence = rje.strSub(an.sequence, r, r, a)
if maxp == 0: # No AA => X
an.sequence = rje.strSub(an.sequence, r, r, "X")
if aafix:
for a in aalist:
if a == an.sequence[r]:
an.ancaap[r][a] = 1.0
else:
an.ancaap[r][a] = 0.0
### Finish ###
self.log.printLog("\r#GASP", "Calculated GASP %s: 100.0%%." % logtxt)
except:
self.log.errorLog("Fatal Error during _gaspProbs().")
raise