B. Schwikowski and M. Vingron. The Deferred Path Heuristic for the Generalized Tree Alignment Problem. To appear in: Proceedings of the First Annual International Conference on Computational Molecular Biology, ACM 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....structure where a given sequence is an internal node. Sometimes, it is unacceptable. Schwikowski and Vingron give a method that combines clustering algorithms and Hein s sequence graph method. The produced solutions contain biologically reasonable trees and keep the guaranteed performance ratio. [67]. 4.4 Fixed topology history alignment with recombination Multigene families, viruses, and alleles from within populations experience recombinations [34, 35, 47, 71] When recombination happens, the ancestral material on the present sequence s 1 is located on two sequences s 2 and s 3 . s 2 and ....

B. Schwikowski and M. Vingron. The deferred path heuristic for the generalized tree alignment problem, 1st Annual International Conference On Computational Molecular Biology, 1997, pp. 257-266.

....of sequence graphs. Furthermore we will present the application of our program to the study of a set of Alu repeats. Another variant of the algorithm with a guaranteed error bound is based on a description of Generalized Tree Alignment as a Steiner Tree Problem. It is described in detail elsewhere [13]. 2 Sequence Graphs When posed for two individual sequences, the Generalized Tree Alignment problem asks for a most parsimonious explanation of their history since their divergence in evolution. The answer is a series of mutations and indels constituting an evolutionary pathway in the space of ....

....G 0 k then represents not only all sequences on any shortest path between G l and Gm , but also the sequences in S l and Sm . The modified algorithm, called DPH F2, calculates a score that is at most (2 Gamma 2 n ) times the score of the optimal solution. The proof for this can be found in [13]. Computational Complexity. A governing factor in the overall complexity of the algorithms DPH AV and DPH F2 is the size of the sequence graphs generated. In terms of memory needed, a sequence graph P(G 1 ; G 2 ) can in the worst case grow to limiting sizes. This might happen if, e.g. G 1 ....

B. Schwikowski and M. Vingron. The Deferred Path Heuristic for the Generalized Tree Alignment Problem. To appear in: Proceedings of the First Annual International Conference on Computational Molecular Biology, ACM 1997.

.... for all sequences s between S(G 1 ) and S(G 2 ) Since there is a gap initiation penalty a 0 there is the aforementioned interdependence between different gap columns 9 (alignment columns with gap symbols) that makes the treatment more complicated than in the case a = 0, which is treated in (Schwikowski and Vingron 1997). The sequences s between S(G 1 ) and S(G 2 ) will be represented as paths through a new directed graph P 0 = P 0 (G 1 ; G 2 ) P 0 will have the same vertex set as A . Its edge set is defined by Table 1: Whenever the labels of a path in A match a pattern, a corresponding edge in P 0 is ....

Schwikowski, B., and Vingron, M. 1997. The deferred path heuristic for the generalized tree alignment problem. In: Proceedings of the 1st Annual International Conference on Computational Molecular Biology , 257--266, ACM Press, Santa Fe, New Mexico, USA.

....perform is a minimization or maximization, one convertible into the other by negating signs. In this paper we assume a minimization framework, keeping in mind that maximization schemes can be accommodated after trivial modification. Previous methods that use the subtree score to select candidates [18, 21, 20] construct candidate sequences by considering letters from optimal pairwise alignments of two descendant sequences. This procedure only finds all optimal candidates when the mutation score function w satisfies min Gamma w(x; x) w(z; z) Delta w(x; z) w(x; y) w(y; z) for all x; z 2 ....

....score of a sequence s at u, according to Lemma 1. When the candidates at u are all sequences with a near optimal subtree score f u (s) however, the size of the candidate sets can be expected to grow considerably, even when only sequences with an optimal subtree score are chosen as candidates [21]. This means that enumerative approaches to calculating near optimal subtree scores, conceivably similar to previous approaches for pairwise alignment [27, 28] would not be feasible. We require a data structure to calculate and represent a set of candidate sequences with their subtree scores much ....

[Article contains additional citation context not shown here]

B. Schwikowski and M. Vingron, The deferred path heuristic for the generalized tree alignment problem, in Proceedings of the 1st Annual International Conference on Computational Molecular Biology, ACM Press (1997), pp. 257--266.

....of its associated edge labels, a candidate sequence, and the union of all paths represents a whole candidate set. The algorithm to calculate sequence graphs for ancestral nodes was later refined and it was shown that the sequence graph indeed represents the sequences between the candidate sets [20]. Since the subtree score is identical for all candidates of a candidate set, it needs not to be represented for each individual sequence. 2.4 Two Essential Improvements In this paper we remove two limitations of previous methods for the case of variable alignments. Our method enables the use of ....

....perform is a minimization or maximization, one convertible into the other by negating signs. In this paper we assume a minimization framework, keeping in mind that maximization schemes can be accommodated after trivial modification. Previous methods that use the subtree score to select candidates [18, 21, 20] construct candidate sequences by considering letters from optimal pairwise alignments of two descendant sequences. This procedure only finds all optimal candidates when the mutation score function w satisfies min Gamma w(x; x) w(z; z) Delta w(x; z) w(x; y) w(y; z) for all x; z 2 ....

[Article contains additional citation context not shown here]

B. Schwikowski and M. Vingron, The deferred path heuristic for the generalized tree alignment problem, J. Comp. Biol., 4, 3 (1997), 415--431.

....a tree given (more accurately, he calculates one based on the pairwise distances using clustering) As he aligns along the tree he represents clusters by sequence graphs. When assigning sequences to inner nodes based on the sequence graph he achieves an effect that may be termed coalescement. In [31] sequence graphs are applied for the design of an approximation algorithm for generalized tree alignment. This is based on a variant of the MST heuristic where the assignment of ancestral sequences is done by selecting from a sequence graph after the topology of the tree was derived. Coalescement ....

B. Schwikowski and M. Vingron. The deferred path heuristic for the generalized tree alignment problem. In RECOMB Proceedings, pages 257--266. ACM press, New York, 1997.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC