MicroRNA target prediction by constraint programming

Official URL: http://risc01.sabanciuniv.edu/record=b1169920 (Table of Contents)

MicroRNAs (miRNAs) are small regulatory RNAs of about 22 nucleotide long sequences that perform important functions such as larval development switches, cell proliferation and differentiation, apoptosis, fat metabolism, control of leaf and flower development. MicroRNA sequences are highly conserved across even unrelated species, a fact which suggests a key role in the evolutionary development. MicroRNAs are transcribed in the nucleus and perform their functions in the cytoplasm by binding to the complementary target mRNAs. MicroRNAs modulate gene expression either by suppressing translation or by mRNA cleavage and degradation. Plant microRNAs bind to their target mRNA on the coding region, almost perfectly, and perform their function by the cleavage of the mRNA, while animal microRNAs, bind imperfectly to their target mRNA, on the 3’ UTR region, and perform their functions by suppressing translation. MicroRNAs are discovered by both mutational studies and by computational methods. Hundreds of microRNAs have been cloned and sequenced in several organisms including humans, but to date, only few of them have known functions. The experimental techniques to understand the functions of miRNAs are time consuming and expensive which makes computational methods necessary. The identification of targets of plant microRNAs is straightforward due to near-perfect binding, but the imperfect binding of animal miRNAs to target mRNAs makes the computational target prediction rather difficult. In this thesis a new method is proposed for microRNA target prediction in animals using Constraint Logic Programming. With the established method a package micTar was developed to identify targets in Drosophila genome.