Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10351945 | Computers in Biology and Medicine | 2005 | 17 Pages |
Abstract
This work applies two recently formulated quantities, strongly correlated with the coding character of a sequence, as an additional “module” on GeneMark, in a three-criterial method. The difference in the statistical approaches implicated by the methods combined here, is expected to contribute to an efficient assignment of functionality to unannotated genomic sequences. The developed combined algorithm is used to fractionalize a collection of GeneMark-predicted exons into sub-collections of different expectation to be coding. A further modification of the algorithm allows for the assignment of an improved estimation of the probability to be coding, to GeneMark-predicted exons. This is on the basis of a suitable training set of GeneMark-predicted exons of known functionality.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science Applications
Authors
Yannis Almirantis, Christoforos Nikolaou,