UBC Theses and Dissertations
Comparative sequence analysis : prediction of consensus Xist RNA secondary structure Chan, Viann Wendy
The typical female mammal has two X chromosomes. However, only one copy remains active, while the other copy is silenced. The Xist gene has been found to remain active on the silenced X chromosome. Other lab experiments have determined that the Xist RNA plays a significant role in the process of X - chromosome inactivation. Our goal in this thesis is to predict a possible consensus structure for the Xist RNA using comparative sequence analysis. Since the Xist RNA structure is not known, we evaluate our approaches on sequences with known structures. There are three main phases for this approach: (i) aligning multiple sequences, (ii) finding stems, and (iii) building the structure. Phases (ii) and (iii) are the main focus of this research. We used the concept of mutual information content to detect potential stems containing mutations. We found potential conserved stems by searching for sequences of complementary paired bases. The stems were then used to find the consensus structure. We investigated two general approaches for building the consensus structure, greedy and randomized methods. We tested these methods on data comprising the following types of RNA: tRNA, 16S rRNA, 23S rRNA, and Xist RNA. We found that our methods accurately predicted the consensus structure for tRNA. However, when we analyzed data sets with longer sequences, the accuracy of our predictions decreased. Of the methods we tested, we found that one of our greedy approaches performed better than the others on all the data sets. We predicted several possible structures for Xist from different alignments. While the overall accuracy of our predicted structures is likely to be low, some of the substructures predicted may give us clues about the true consensus structure.
Item Citations and Data