Affiliations: Department of Computer Science, Ben-Gurion University,
Beer Sheva 84105, Israel. E-mail: {bergig, dbarash,
klara}@cs.bgu.ac.il | Department of Biochemistry, New York University
Medical Center, NY 10016, USA. E-mail: [email protected]
Note: [] Corresponding author
Abstract: Traditional sequence-based search methods such as BLAST and FASTA
can be used to identify sequence similarities. Recently, there is a growing
interest in performing RNA shape similarity searches inside selected genes to
locate RNA structure motifs that are known to possess functionally important
roles. For example, in the newly discovered RNA genetic control elements called
"riboswitches", the box domain is known to be highly conserved among various
bacterial species in both its nucleotide composition and shape. However, in
non-bacterial species, shape conservation is likely to become more important
than sequence conservation when searching for riboswitch patterns. For this
purpose, we present an approach tailored for detecting RNA shape similarities.
We extend the Structure to String (STR^{2}) method that was
initially proposed to locate shape similarities in proteins to identify
predicted secondary structures of RNAs. The STR^{2} for RNAs
is a translation of a secondary structure to a string of characters, after
which known sequence-based search algorithms with an efficient implementation
are being used. We validate that the STR^{2} succeeds to
locate G-box riboswitches in prokaryotes, as expected. Subsequently we show
running examples when attempting to detect G-box riboswitch candidates in
eukaryotes.
