Dynamic De-novo prediction of microRNAs associated with cell conditions: A search pruned by expression

Chaya Ben Zaken Zilberstein, Michal Ziv-Ukelson

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Biological background: Plant microRNAs (miRNAs) are short RNA sequences that bind to target genes (niRNAs) and change their expression levels by redirecting their stabilities and marking them for cleavage, In Arabidopsis thaliana, microRNAs have been shown to regulate development and are believed to impact expression both under many conditions, such as stress and stimuli, us well as in various tissue types. Methods: mirXdeNovo is a novel prototype tool for the de-novo prediction of microRNAs associated with a given cell condition. The work of mirXdeNovo is composed of two off-line preprocessing stages, which are executed only once per genome in the database, and a dynamic online main stage, which is executed again and again for each newly obtained expression profile. During the preprocessing stages, a set of candidate microRNAs is computed for the genome of interest and then each microRNA is associated with a set of mRNAs which are its predicted targets. Then, during the main stage, given a newly obtained cell condition represented by a vector describing the expression level of each of the genes under this condition, the tool will efficiently compute the subset of microRNA candidates which are predicted to be active under this condition. The efficiency of the main stage is based in a novel branch-and-bound search of a tree constructed over the microRNA candidates and annotated with the corresponding predicted targets. This search exploits the monotonicity of the target prediction decision with respect to microRNA prefix size in order to apply an efficient yet admissible pruning. Our testing indicates that this paining results in a substantial speed up over the naive search. Biological Results: We employed mirXdeNovo to conduct a study, using the plant Arabidopsis thaliana as our model organism and the subject of our "hunt for microRNAs". During the preprocessing stage, 2000 microRNA precursor candidates were extracted from the genome. Our study included the 3'UTRs of 5800 mRNAs. 380 different conditions were analyzed including various tissues and hormonal treatments. This led to the discovery of some interesting and statistically significant newly predicted microRNAs, annotated with their potential condition of activity.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages13-26
Number of pages14
DOIs
StatePublished - 1 Dec 2005
Externally publishedYes
Event5th International Workshop on Algorithms in Bioinformatics, WABI 2005 - Mallorca, Spain
Duration: 3 Oct 20056 Oct 2005

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3692 LNBI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference5th International Workshop on Algorithms in Bioinformatics, WABI 2005
Country/TerritorySpain
CityMallorca
Period3/10/056/10/05

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Dynamic De-novo prediction of microRNAs associated with cell conditions: A search pruned by expression'. Together they form a unique fingerprint.

Cite this