A parallel algorithm for the extraction of structured motifs

Carvalho, Alexandra M.; Oliveira, Arlindo L.; Freitas, Ana T.; Sagot, Marie‐France

doi:10.1145/967900.967932

Cited by 25 publications

(34 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From an algorithmic point of view, and in particular in what concerns parallel algorithms, we shall apply the parallelization technique proposed in [22] to the algorithm at hand. This approach will allow the use of even larger datasets which is a major limitation of current solutions.…”

Section: Resultsmentioning

confidence: 99%

Efficient Extraction of Structured Motifs Using Box-Links

Carvalho¹,

Freitas²,

Oliveira³

et al. 2004

String Processing and Information Retrieval

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a new algorithm for the extraction of repeated motifs that may represent binding-site consensi in genomic sequences. In particular, the algorithm extracts structured motifs, which we define as a collection of highly conserved motifs with pre-specified sizes and spacings between them. This type of motifs is highly relevant in the search for gene regulatory mechanisms since promoter models can be effectively represented by structured motifs.The algorithm uses factor trees, a variation of suffix trees, and a new data structure, called box-links, to store the information about conserved regions that repeat often in the dataset sequences. The complexity analysis shows a gain over previous algorithms that is exponential on the spacings between boxes.The application of a prototype implementation of this algorithm to biologically relevant datasets shows the ability of the method to extract relevant consensi. The experimental results also show that this algorithm is much faster than existing ones, sometimes by more than two orders of magnitude.

show abstract

Section: Resultsmentioning

confidence: 99%

Efficient Extraction of Structured Motifs Using Box-Links

Carvalho¹,

Freitas²,

Oliveira³

et al. 2004

String Processing and Information Retrieval

Self Cite

View full text Add to dashboard Cite

show abstract

“…It works by constructing a generalized suffix tree using the input sequences and then finding the motifs (or spelling the models) using this generalized suffix tree. Following that many modifications and extensions were proposed to improve the performance [8] [9] [10]. In this paper we adapt the original SPELLER algorithm for efficient implementation on multicore and GPU.…”

Section: The Core Suffix Tree Algorithmmentioning

confidence: 99%

“…SPELLER is a tree based algorithm and suffix tree by nature is very unbalanced. A previous attempt to parallelize SPELLER can be found in [10]. In [10] the count of the residues is used as the basis for distributing the load among multiple nodes.…”

Section: Adapting Speller On Multicorementioning

confidence: 99%

“…MITRA is more efficient than SPELLER in terms of both memory and space. SMILE [9], PSMILE [10], RISOTTO [8] are extensions to the SPELLER algorithm, RISOTTO being the most recent approach. It uses maximum extensibility to efficiently spell the motifs.…”

Section: Introductionmentioning

confidence: 99%

“…There have not been many attempts to propose a parallel approach based on suffix tree. To the best of our knowledge, PSMILE [10] is the only parallel suffix tree based approach that has been proposed. It can be seen from PSMILE that it is not straightforward to parallelize a suffix tree based approach as it is difficult to distribute the work equally among all the processors.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

High performance implementation of planted motif problem using suffix trees

Dasari

Ranjan

Zubair

2011

2011 International Conference on High Performance Computing &Amp; Simulation

View full text Add to dashboard Cite

show abstract

High‐performance implementation of planted motif problem on multicore and GPU

Dasari

Ranjan

Zubair

2012

Concurrency and Computation

View full text Add to dashboard Cite

In this paper, we present an efficient, easily parallelizable approach to solve planted motif problem (PMP). PMP is a well-studied problem in computational biology. It is useful in developing methods for finding transcription factor binding sites, classifying sequences, and building phylogenetic trees. Many approaches to solve PMP can be found in the literature. But the problem with those approaches is that they are difficult to parallelize as they have been designed for serial computers. In this paper, we propose a simple, easily parallelizable enumeration-based approach called BitBased. As with most other enumeration-based approaches that have been proposed to solve PMP, BitBased is also limited by memory for solving largesized problems. To overcome this limitation, we propose various modifications, which not only reduce the memory requirement but also improve the performance of the approach. We have implemented our approach on multicore and GPU devices. We found that BitBased outperforms all the approaches proposed to solve PMP so far. BitBased is able to solve the (21, 8) instance, which was not previously reported as solved in the literature.Many approaches have been previously proposed to solve the PMP. These approaches can be classified into two categories: heuristic and exact. Heuristic algorithms are very popular, but they are not guaranteed to find the correct answer always. CONSENSUS [4], WINNOWER [2], and Random Projections [5] are some approaches that fall under this category. Exact algorithms, on the other hand, always produce the correct answer. These algorithms are also referred to as exhaustive IMPLEMENTATION OF PMP ON MULTICORE AND GPU 1341 enumeration algorithms. SPELLER [6], MITRA [7], PMSprune [8], Voting [9], and RISOTTO [10] are some approaches that fall under this category. These algorithms can further be classified into pattern-driven and sample-driven approaches. Pattern-driven approaches search all the possible j †j l l-mers to find the motifs where † D fA, C , G, T g. These algorithms have the time complexity of O.j †j l /. These algorithms are therefore only suitable for smaller values of l and perform prohibitively poorer for larger values of l. Sample-driven approaches, on the other hand, enumerate the l-mers in the input sequences to find the motifs. Sample-driven approaches are often limited by space requirement. BitBased approach is an exact sample-driven approach. Like other sampledriven approaches, BitBased also has the disadvantage of high space complexity. But we propose various techniques to reduce the space requirement. BitBased can be easily modified to work with any amount of memory available but with additional time.Sagot [6] introduced a suffix tree-based algorithm for solving PMP. This algorithm starts by building a generalized suffix tree for all the input sequences and uses this tree to 'spell' all the motifs. This algorithm is very efficient in terms of space. MITRA uses a variation of suffix tree called Mismatch trees. MITRA works by splitting all the possible pa...

show abstract

A parallel algorithm for the extraction of structured motifs

Cited by 25 publications

References 13 publications

Efficient Extraction of Structured Motifs Using Box-Links

Efficient Extraction of Structured Motifs Using Box-Links

High performance implementation of planted motif problem using suffix trees

High‐performance implementation of planted motif problem on multicore and GPU

Contact Info

Product

Resources

About