def alignment(self):
"""Make self into an alignment, and return it.
If all the sequences are the same length and type, then self,
a sequenceList, could be an Alignment. This method generates
an Alignment instance, runs the Alignment method
checkLengthsAndTypes(), and returns the Alignment.
If you feed p4 a fasta sequence, it makes SequenceList object,
and runs this method on it. If it works then p4 puts the
Alignment object in var.alignments, and if not it puts the
SequenceList object in var.sequenceLists.
It is possible that p4 might think that some short sequences
are DNA when they are really protein. In that case it will
fail to make an alignment, because it will fail the types
check. So what you can do is something like this::
sl = var.sequenceLists[0]
for s in sl.sequences:
s.dataType = 'protein'
a = sl.alignment()
"""
from alignment import Alignment
a = Alignment()
a.fName = self.fName
import copy
a.sequences = copy.deepcopy(self.sequences) # self will be deleted
a.fName = self.fName
a.checkLengthsAndTypes()
return a
def alignment(self):
"""Make self into an alignment, and return it.
If all the sequences are the same length and type, then self,
a sequenceList, could be an Alignment. This method generates
an Alignment instance, runs the Alignment method
checkLengthsAndTypes(), and returns the Alignment.
If you feed p4 a fasta sequence, it makes SequenceList object,
and runs this method on it. If it works then p4 puts the
Alignment object in var.alignments, and if not it puts the
SequenceList object in var.sequenceLists.
It is possible that p4 might think that some short sequences
are DNA when they are really protein. In that case it will
fail to make an alignment, because it will fail the types
check. So what you can do is something like this::
sl = var.sequenceLists[0]
for s in sl.sequences:
s.dataType = 'protein'
a = sl.alignment()
"""
from alignment import Alignment
a = Alignment()
a.fName = self.fName
import copy
a.sequences = copy.deepcopy(self.sequences) # self will be deleted
a.fName = self.fName
a.checkLengthsAndTypes()
return a
def calcUnconstrainedLogLikelihood1(self):
"""Calculate likelihood under the multinomial model.
This calculates the unconstrained (multinomial) log like
without regard to character partitions. The result is placed
in the data variable unconstrainedLogLikelihood. If there is
more than one partition, it makes a new temporary alignment
and puts all the sequences in one part in that alignment. So
it ultimately only works on one data partition. If there is
more than one alignment, there is possibly more than one
datatype, and so this method will refuse to do it. Note that
the unconstrained log like of the combined data is not the sum
of the unconstrained log likes of the separate partitions.
See also calcUnconstrainedLogLikelihood2
"""
if len(self.alignments) > 1:
gm = ["Data.calcUnconstrainedLogLikelihood()"]
gm.append("This method is not implemented for more than one alignment.")
raise P4Error(gm)
if self.nParts == 1: # no problem
self.unconstrainedLogLikelihood = pf.getUnconstrainedLogLike(self.parts[0].cPart)
else:
a = self.alignments[0]
import copy
newAlig = Alignment()
newAlig.dataType = a.dataType
newAlig.symbols = a.symbols
newAlig.dim = a.dim
newAlig.equates = a.equates
newAlig.taxNames = a.taxNames
for s in a.sequences:
newAlig.sequences.append(copy.deepcopy(s))
newAlig.checkLengthsAndTypes()
newAlig._initParts()
# newAlig.dump()
self.unconstrainedLogLikelihood = pf.getUnconstrainedLogLike(newAlig.parts[0].cPart)
del (newAlig)
