Research article
Shape based indexing for faster search of RNA family databases
-
* Corresponding author: Robert Giegerich robert@techfak.uni-bielefeld.de
Faculty of Technology, Bielefeld University, 33615 Bielefeld, Germany
BMC Bioinformatics 2008, 9:131 doi:10.1186/1471-2105-9-131
Published: 29 February 2008Additional files
Additional file 1:
Effects of skipping "difficult" families on sensitivity and filtration ratio. The test- and training sets are constructed as above, but this time we choose up to 1,000 sequences for each family instead of four. Random- and gene- testsets are not considered, because we focus on the changes of the sensitivity. RNAsifter is set to kfamily = 3, kquery = 5, ε = 0.4. One has to read the rows in a accumulative fashion. In the first row no family is skipped, in the second row family RF00017 is omitted, the third row omits families RF00017 and RF00230 and so on. Note that the testset shrinks, because the sequences of a skipped family are also removed from the set.
Format: PDF Size: 40KB Download file
This file can be viewed with: Adobe Acrobat Reader
