Search PubMed⌕ Search

Biomedical subjects

Dan Gusfield

Publications and source records attributed to Dan Gusfield.

8 recordsLinked to original sources
PubMed2006

A linear-time algorithm for the perfect phylogeny haplotyping (PPH) problem.

Since the introduction of the Perfect Phylogeny Haplotyping (PPH) Problem in RECOMB 2002 (Gusfield, 2002), the problem of finding a linear-time (deterministic, worst-case) solution for it has remained open, despite broad interest in the PPH problem and a series of papers on various aspects of it. In this paper, we solve the open problem, giving a practical, deterministic linear-time algorithm based on a simple data structure and simple operations on it. The method is straightforward to program and has been fully implemented. Simulations show that it is much faster in practice than prior nonlinear methods. The value of a linear-time solution to the PPH problem is partly conceptual and partly for use in the inner loop of algorithms for more complex problems, where the PPH problem must be solved repeatedly.

Algorithms↗
PubMed2005

Efficient computation of close lower and upper bounds on the minimum number of recombinations in biological sequence evolution.

MOTIVATION: We are interested in studying the evolution of DNA single nucleotide polymorphism sequences which have undergone (meiotic) recombination. For a given set of sequences, computing the minimum number of recombinations needed to explain the sequences (with one mutation per site) is a standard question of interest, but it has been shown to be NP-hard, and previous algorithms that compute it exactly work either only on very small datasets or on problems with special structure. RESULTS: In this paper, we present efficient, practical methods for computing both upper and lower bounds on the minimum number of needed recombinations, and for constructing evolutionary histories that explain the input sequences. We study in detail the efficiency and accuracy of these algorithms on both simulated and real data sets. The algorithms produce very close upper and lower bounds, which match exactly in a surprisingly wide range of data. Thus, with the use of new, very effective lower bounding methods and an efficient algorithm for computing upper bounds, this approach allows the efficient, exact computation of the minimum number of needed recombinations, with high frequency in a large range of data. When upper and lower bounds match, evolutionary histories found by our algorithm correspond to the most parsimonious histories. AVAILABILITY: HapBound and SHRUB, programs implementing the new algorithms discussed in this paper, are available at http://wwwcsif.cs.ucdavis.edu/~gusfield/lu.html

Algorithms↗
PubMed2004

A note on efficient computation of haplotypes via perfect phylogeny.

The problem of inferring haplotype phase from a population of genotypes has received a lot of attention recently. This is partly due to the observation that there are many regions on human genomic DNA where genetic recombination is rare (Helmuth, 2001; Daly et al., 2001; Stephens et al., 2001; Friss et al., 2001). A Haplotype Map project has been announced by NIH to identify and characterize populations in terms of these haplotypes. Recently, Gusfield introduced the perfect phylogeny haplotyping problem, as an algorithmic implication of the no-recombination in long blocks observation, together with the standard population-genetic assumption of infinite sites. Gusfield's solution based on matroid theory was followed by direct theta(nm2) solutions that use simpler techniques (Bafna et al., 2003; Eskin et al., 2003), and also bound the number of solutions to the PPH problem. In this short note, we address two questions that were left open. First, can the algorithms of Bafna et al. (2003) and Eskin et al. (2003) be sped-up to O(nm + m2) time, which would imply an O(nm) time-bound for the PPH problem? Second, if there are multiple solutions, can we find one that is most parsimonious in terms of the number of distinct haplotypes. We give reductions that suggests that the answer to both questions is "no." For the first problem, we show that computing the output of the first step (in either method) is equivalent to Boolean matrix multiplication. Therefore, the best bound we can presently achieve is O(nm(omega-1)), where omega < or = 2.52 is the exponent of matrix multiplication. Thus, any linear time solution to the PPH problem likely requires a different approach. For the second problem of computing a PPH solution that minimizes the number of distinct haplotypes, we show that the problem is NP-hard using a reduction from Vertex Cover (Garey and Johnson, 1979).

Computational Biology↗
PubMed2004

Optimal, efficient reconstruction of phylogenetic networks with constrained recombination.

A phylogenetic network is a generalization of a phylogenetic tree, allowing structural properties that are not tree-like. In a seminal paper, Wang et al.(1) studied the problem of constructing a phylogenetic network, allowing recombination between sequences, with the constraint that the resulting cycles must be disjoint. We call such a phylogenetic network a "galled-tree". They gave a polynomial-time algorithm that was intended to determine whether or not a set of sequences could be generated on galled-tree. Unfortunately, the algorithm by Wang et al.(1) is incomplete and does not constitute a necessary test for the existence of a galled-tree for the data. In this paper, we completely solve the problem. Moreover, we prove that if there is a galled-tree, then the one produced by our algorithm minimizes the number of recombinations over all phylogenetic networks for the data, even allowing multiple-crossover recombinations. We also prove that when there is a galled-tree for the data, the galled-tree minimizing the number of recombinations is "essentially unique". We also note two additional results: first, any set of sequences that can be derived on a galled tree can be derived on a true tree (without recombination cycles), where at most one back mutation per site is allowed; second, the site compatibility problem (which is NP-hard in general) can be solved in polynomial time for any set of sequences that can be derived on a galled tree. Perhaps more important than the specific results about galled-trees, we introduce an approach that can be used to study recombination in general phylogenetic networks. This paper greatly extends the conference version that appears in an earlier work.(8) PowerPoint slides of the conference talk can be found at our website.(7).

Algorithms↗
PubMed2003

On the complexity of fundamental computational problems in pedigree analysis.

Pedigree analysis is a central component of many current efforts to locate genes that contribute to diseases or to valuable traits. The analysis usually involves solving one of two very computation-intense problems. We analyze the complexity of these two problems. Surprisingly, we show that both problems are NP-hard even for pedigrees that contain no inbreeding loops.

Computational Biology↗
PubMed2003

Haplotyping as perfect phylogeny: a direct approach.

A full haplotype map of the human genome will prove extremely valuable as it will be used in large-scale screens of populations to associate specific haplotypes with specific complex genetic-influenced diseases. A haplotype map project has been announced by NIH. The biological key to that project is the surprising fact that some human genomic DNA can be partitioned into long blocks where genetic recombination has been rare, leading to strikingly fewer distinct haplotypes in the population than previously expected (Helmuth, 2001; Daly et al., 2001; Stephens et al., 2001; Friss et al., 2001). In this paper we explore the algorithmic implications of the no-recombination in long blocks observation, for the problem of inferring haplotypes in populations. This assumption, together with the standard population-genetic assumption of infinite sites, motivates a model of haplotype evolution where the haplotypes in a population are assumed to evolve along a coalescent, which as a rooted tree is a perfect phylogeny. We consider the following algorithmic problem, called the perfect phylogeny haplotyping problem (PPH), which was introduced by Gusfield (2002) - given n genotypes of length m each, does there exist a set of at most 2n haplotypes such that each genotype is generated by a pair of haplotypes from this set, and such that this set can be derived on a perfect phylogeny? The approach taken by Gusfield (2002) to solve this problem reduces it to established, deep results and algorithms from matroid and graph theory. Although that reduction is quite simple and the resulting algorithm nearly optimal in speed, taken as a whole that approach is quite involved, and in particular, challenging to program. Moreover, anyone wishing to fully establish, by reading existing literature, the correctness of the entire algorithm would need to read several deep and difficult papers in graph and matroid theory. However, as stated by Gusfield (2002), many simplifications are possible and the list of "future work" in Gusfield (2002) began with the task of developing a simpler, more direct, yet still efficient algorithm. This paper accomplishes that goal, for both the rooted and unrooted PPH problems. It establishes a simple, easy-to-program, O(nm(2))-time algorithm that determines whether there is a PPH solution for input genotypes and produces a linear-space data structure to represent all of the solutions. The approach allows complete, self-contained proofs. In addition to algorithmic simplicity, the approach here makes the representation of all solutions more intuitive than in Gusfield (2002), and solves another goal from that paper, namely, to prove a nontrivial upper bound on the number of PPH solutions, showing that that number is vastly smaller than the number of haplotype solutions (each solution being a set of n pairs of haplotypes that can generate the genotypes) when the perfect phylogeny requirement is not imposed.

Algorithms↗
PubMed2003

Efficient reconstruction of phylogenetic networks with constrained recombination.

A phylogenetic network is a generalization of a phylogenetic tree, allowing structural properties that are not tree-like. With the growth of genomic data, much of which does not fit ideal tree models, there is greater need to understand the algorithmics and combinatorics of phylogenetic networks [10, 11]. However, to date, very little has been published on this, with the notable exception of the paper by Wang et al.[12]. Other related papers include [4, 5, 7] We consider the problem introduced in [12], of determining whether the sequences can be derived on a phylogenetic network where the recombination cycles are node disjoint. In this paper, we call such a phylogenetic network a "galled-tree". By more deeply analysing the combinatorial constraints on cycle-disjoint phylogenetic networks, we obtain an efficient algorithm that is guaranteed to be both a necessary and sufficient test for the existence of a galled-tree for the data. If there is a galled-tree, the algorithm constructs one and obtains an implicit representation of all the galled trees for the data, and can create these in linear time for each one. We also note two additional results related to galled trees: first, any set of sequences that can be derived on a galled tree can be derived on a true tree (without recombination cycles), where at most one back mutation is allowed per site; second, the site compatibility problem (which is NP-hard in general) can be solved in linear time for any set of sequences that can be derived on a galled tree. The combinatorial constraints we develop apply (for the most part) to node-disjoint cycles in any phylogenetic network (not just galled-trees), and can be used for example to prove that a given site cannot be on a node-disjoint cycle in any phylogenetic network. Perhaps more important than the specific results about galled-trees, we introduce an approach that can be used to study recombination in phylogenetic networks that go beyond galled-trees.

Computer Simulation↗