Heuristic Reusable Dynamic Programming: Efficient Updates of Local Sequence Alignment

Recomputation of the previously evaluated similarity results between biological sequences becomes inevitable when researchers realize errors in their sequenced data or when the researchers have to compare nearly similar sequences, e.g., in a family of proteins. We present an efficient scheme for upd...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	IEEE/ACM transactions on computational biology and bioinformatics Ročník 6; číslo 4; s. 570 - 582
Hlavní autoři:	Changjin Hong, Tewfik, A.H.
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	United States IEEE 01.10.2009 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:	Algorithms Artificial Intelligence Computational Biology - methods Computer Graphics Dynamic programming Error analysis Error correction Genetic mutations Humans Laboratories minimum spanning tree Pattern Recognition, Automated Polymers Programming Languages Proteins Proteins - chemistry Reproducibility of Results sensitivity analysis sequence alignment Sequence Alignment - methods Sequence Analysis, Protein - methods Sequences Shortest path Signal Processing, Computer-Assisted Software string edit Studies suboptimal paths Switches
ISSN:	1545-5963, 1557-9964, 1557-9964
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Recomputation of the previously evaluated similarity results between biological sequences becomes inevitable when researchers realize errors in their sequenced data or when the researchers have to compare nearly similar sequences, e.g., in a family of proteins. We present an efficient scheme for updating local sequence alignments with an affine gap model. In principle, using the previous matching result between two amino acid sequences, we perform a forward-backward alignment to generate heuristic searching bands which are bounded by a set of suboptimal paths. Given a correctly updated sequence, we initially predict a new score of the alignment path for each contour to select the best candidates among them. Then, we run the Smith-Waterman algorithm in this confined space. Furthermore, our heuristic alignment for an updated sequence shows that it can be further accelerated by using reusable dynamic programming (rDP), our prior work. In this study, we successfully validate "relative node tolerance boundrdquo (RNTB) in the pruned searching space. Furthermore, we improve the computational performance by quantifying the successful RNTB tolerance probability and switch to rDP on perturbation-resilient columns only. In our searching space derived by a threshold value of 90 percent of the optimal alignment score, we find that 98.3 percent of contours contain correctly updated paths. We also find that our method consumes only 25.36 percent of the runtime cost of sparse dynamic programming (sDP) method, and to only 2.55 percent of that of a normal dynamic programming with the Smith-Waterman algorithm.
Bibliografie:	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 ObjectType-Article-2 ObjectType-Feature-1 content type line 23
ISSN:	1545-5963 1557-9964 1557-9964
DOI:	10.1109/TCBB.2009.30