This module contains low-level code for finding the optimal alignment between two motifs.
The central algorithm is called "minshortestoverhangdiff," which searches all alignments
of two motifs to find the one that minimizes the specificied distance/divergence metric.
The space of alignments is restricted by the minimum amount of overlap that is required
between motifs, which is in turn computed on the fly using the "OVLP" function.
 
Distance/divergence metrics include:
    -== To be completed ==-
 
Copyright (2005) Whitehead Institute for Biomedical Research (except as noted below)
All Rights Reserved
 
Author: David Benjamin Gordon

Modules
		TAMO.util.Arith TAMO.MotifTools math os random re sys

Functions
		ENMrange(self, other, Srange, Orange) Utility function:  compute Expected Number of Mismatches NCBpenalty(penalty=0) NLBPO(A, B, thresh) NLPO(A, B, thresh) OVLP lambda A, B, w=2 averagemotifs(motifs, ovlp=2, template=None, DFUNC=<function negcommonbitsrange>, VERBOSE=1, prop='') averagetxts(txts, ovlp=2, VERBOSE=1) centroididx(Dmat) coalesce(motifs, DMIN=0.22) coarsecluster(motifs) computeDmat(motifs, VERBOSE=0, DFUNC=None) diffrange(self, other, Srange, Orange) Utility function: compute diff of entropyrange(self, other, Srange, Orange) Utility function: compute diff of estimate_frequency(motif, k, samples=100000, thresh=0.69999999999999996) etestdiff(t1, t2, OVLP_FCN) genNCBpenalty(penalty=0) main() minshortestoverhangdiff(A, B, minoverlap=6, want_offset=None, DFUNC=None) negcommonbitsrange(self, other, Srange, Orange, PENALTY=0) Utility function: compute common bits in self and other   Common bits are computed using this expression: SUM [ max(PENALTY,min(bitsA,bitsB)) ]            (Penalty usually == 0)   where bitsA = Pb_ij * log(Pa_ij / Q_j)   [i = position, j = base letter] negcommonbitstest(t1, t2, OVLP_FCN=None) negprobproductrange(self, other, Srange, Orange) Utility function: compute sum of products of probabilities   SUM [ probOvlp(A, B, thresh=0.69999999999999996, verbose=None) probOvlpBinomial(A, B, thresh=0.69999999999999996, verbose=None) showdiffXvert(motif, seq, OVLP_FCN=None, DIFF_FCN=None) The funtion converts the sequence to a Motif, computes the D of the best alignment, and prints the alignment that generated that D. snegcommonbitsrange(self, other, Srange, Orange) snegbits -- between 0 and 1? snegprobproductrange(self, other, Srange, Orange) sneg -- between 0 and 1? testdiff(t1, t2, OVLP_FCN=None, DIFF_FCN=None)

Data
		ACGT = ['A', 'C', 'G', 'T'] DFUNC = None rcmemo = {}