Darwin Help

Back to Index

DynProgScore

Function DynProgScore - compute the forward phase of sequence alignment

Calling Sequence  DynProgScore(seq1,seq2,dm,modif)
Parameters
NameTypeDescription

seq1 {ProbSeq,string}first sequence to be aligned
seq2 {ProbSeq,string}second sequence to be aligned
dm {DayMatrix,list(DayMatrix)}Dayhoff matrix to use for the alignment
modif {string,set(string)}specification of alignment
Return Type  {[Score:numeric],[score:numeric, from1..to1, from2..to2]}
Synopsis Computes the optimal cost of the alignment between seq1 and seq2 using the Dayhoff matrix dm, a specified alignment mode and a specified deletion cost model. It returns a triplet: [ Score, from1..to1, from2..to2 ] or [ Score ] where Score is the optimal score of the alignment and seq1[from1..to1] and seq2[from2..to2] are the selected portions of the sequences to align. seq1 and seq2 can be either peptide sequences, nucleotide sequences or probabilistic sequences, ProbSeq(). Modif is a set of strings which have the following meanings:

For the alignment type, one of the following can be specified:


Local - (default) a local alignment, the subsequences of seq1 and seq2 which give the highest score.


Global - a global alignment, the entire seq1 is matched against the entire seq2.


CFE - cost-free ends, the entire seq1 is matched against seq2, but one deletion at the ends is not penalized.


CFEright - cost-free ends, the entire seq1 is matched against seq2, but one deletion at the right end is not penalized.


Shake - align seq1 and seq2 up to the point where the maximum score happens. Then do the same backwards and forwards until no improvements of the score happen.


MinLength(k) - align seq1 and seq2 as in a Local alignment (starting anyplace, ending anyplace) but at least k amino acids of each sequence are aligned. I.e. the minimum of the aligned lengths is k or larger.


For the deletion cost model, one of the following can be specified:


Affine - (default) deletion cost of a gap of length k is FixedDel + IncDel*(k-1). The values for FixedDel and IncDel are taken from the Dayhoff matrix dm.


LogDel - logarithmic deletion cost, the cost of a gap of length k is DelFixedLog + DelLog*log(k). The values for DelFixedLog and DelLog are taken from the Dayhoff matrix dm.


For the type of result, any combination of the following can be specified:


JustScore - only the score is computed, and the locations of the match are not returned (this makes the algorithm run faster for Local and CFE).


NoSelf - compute an alignment where matches of a position with itself are disallowed. This is relevant when aligning a sequence with itself with the purpose of discovering repeated motifs.

Examples
> DB := ReadDb('/home/darwin/DB/SwissProt.Z'):;
Peptide file(/home/darwin/DB/SP45.0/SwissProt45.0(169638448), 163235
 entries, 59631787 aminoacids)
> DynProgScore(AC(P00083),AC(P00091),DM,Local);
[177.7799, 14..92, 19..97]
> DynProgScore(AC(P00083),AC(P00091),DM,Global);
[144.4751, 1..127, 1..139]
> DynProgScore(AC(P00083),AC(P00091),DM,{CFE,JustScore});
[174.0188]
> DynProgScore('ADEFGHIKSDEFGHLK','ADEFGHIKSDEFGHLK',DM,NoSelf);
[75.0720, 1..16, 1..16]
See also Align,   Alignment,   CreateDayMatrices,   MAlign