Accurate details in 3D structures of RNA molecules are important for understanding RNA function, which can in turn help us understand biological systems, develop new medicines, and improve human health. One issue in RNA 3D structure analysis is that the structures obtained from biological experiments often contain errors and need to be corrected. The main reason for the errors is because RNA 3D structures are highly complex. While there are existing automatic tools for obtaining protein 3D structures from experimental data, such tools are not yet available for RNAs.
Previously the PI developed a program called RNA Backbone Correction (RNABC) that uses geometric algorithms and robotics to search for error-free RNA structures. While RNABC has been used to correct structural errors in existing RNA structures and has been integrated into the MolProbity web service for RNA structure validation, further advancement is needed. Research shows that RNABC corrects errors in 40% to 80% of RNA structures tested.
This project develops a new and improved computer program that scientists can use to correct structural errors and obtain accurate details of RNA 3D structures. The goal is to correct errors in over 80% of RNA structures and maintain the same running time. The new program will combine methods in machine learning, data mining, robotics and numerical analysis to search for error-free RNA structures. The project will advance algorithmic research in multiple ways and help us better understand the details of RNA structures. This project provides research opportunities to students interested in computer science, mathematics, and biology, and helps educate the next generation of scientists.
Accurate details in 3D structures of RNA molecules are important for understanding RNA function, which can in turn help us understand biological systems, develop new medicines, and improve human health. One issue in RNA 3D structure analysis is that the structures obtained from biological experiments often contain errors and need to be corrected. The main reason for the errors is because RNA 3D structures are highly complex. While there are existing automatic tools for obtaining protein 3D structures from experimental data, such tools are not yet available for RNAs. Previously the PI developed a program called RNA Backbone Correction (RNABC) that uses geometric algorithms and robotics to search for error-free RNA structures. While RNABC has been used to correct structural errors in existing RNA structures and has been integrated into the MolProbity web service for RNA structure validation, further advancement is needed. Research shows that RNABC corrects errors in 40% to 80% of RNA structures tested. This project develops a new and improved computer program that scientists can use to correct structural errors and obtain accurate details of RNA 3D structures, with the goal of correcting errors in over 80% of RNA structures and maintaining the same running time. The new program combines methods in machine learning, data mining, robotics and numerical analysis to search for error-free RNA structures, and advances algorithmic research in multiple ways and helps us better understand the details of RNA structures. The project helped train 5 undergraduate students in bioinformatics research as well as algorithmic thinking. Good progress has been made in developing the new version of RNABC. Although the PI has left NNU in 2013, the PI has already made some updates on the RNABC software as well as some optimizations in k-nearest neighbors and k-means clustering algorithms and will continue to improve the software in the future.
