Improvements in the massively parallel genetic algorithm (GA) for RNA structure prediction are continuing. As well as the porting of the algorithm to a MIMD CRAY T3E. The addition of the annealing mutation operator has substantially improved the results and performance of the algorithm when compared to more standard predictive algorithms such as the dynamic programming algorithm. Pseudoknot prediction has also been included resulting in the ability to predict tertiary interactions. Work has also begun with apparent initial success of including heuristics which promote the formation of short range interactions in the folding process. This again appears to boost the accuracy of structural predictions when comparing results with phylogenetic data. The GA has also been partially integrated into RNA2D3D so that real time folding and the presentation of the three-dimensional RNA structures on an SGI workstation can now be followed. In addition, STRUCTURELAB, the hetergeneous RNA analysis workbench has been further enhanced in conjunction with RNA2D3D to produce the atomic coordinates of RNA structures along with the visualization of these structures. Also, further optimizations and improvements have been made to greatly enhance its capabilities and performance. Porting has also been done to other SGI workstations. We have applied the above techniques to compare secondary structure of the 5' non-coding region of poliovirus 3 RNA derived from the genetic algorithm with the standard Skinner model which demonstrated improved results over conventional techniques. Also, these techniques have been applied to understand the genetics behind, and mechanisms of, the expression of the viral cardiovirulence phenotype of the CVBs which has demonstrated novel structural elements. We have continued to study the dimerization site of HIV-1 RNA. Based on further experiments which include gel shifts, ultracentrifugation, Rnase mapping studies and computational approaches, we have characterized the binding affinity of the nucleocapsid protein NCp7 to the so-called stem-loop 1 and stem-loop 2 RNA structures. We have recently found much higher affinity to stem loop 2 through ultracentrifugation.
Showing the most recent 10 out of 16 publications
