Predicting Large RNA-Like Topologies by a Knowledge-Based Clustering Approach

Naoto Baba, Shereef Elmetwaly, Namhee Kim, Tamar Schlick

Research output: Contribution to journalArticle

Abstract

An analysis and expansion of our resource for classifying, predicting, and designing RNA structures, RAG (RNA-As-Graphs), is presented, with the goal of understanding features of RNA-like and non-RNA-like motifs and exploiting this information for RNA design. RAG was first reported in 2004 for cataloging RNA secondary structure motifs using graph representations. In 2011, the RAG resource was updated with the increased availability of RNA structures and was improved by utilities for analyzing RNA structures, including substructuring and search tools. We also classified RNA structures as graphs up to 10 vertices (~ 200 nucleotides) into three classes: existing, RNA-like, and non-RNA-like using clustering approaches. Here, we focus on the tree graphs and evaluate the newly founded RNAs since 2011, which also support our refined predictions of RNA-like motifs. We expand the RAG resource for large tree graphs up to 13 vertices (~ 260 nucleotides), thereby cataloging more than 10 times as many secondary structures. We apply clustering algorithms based on features of RNA secondary structures translated from known tertiary structures to suggest which hypothetical large RNA motifs can be considered "RNA-like". The results by the PAM (Partitioning Around Medoids) approach, in particular, reveal good accuracy, with small error for the largest cases. The RAG update here up to 13 vertices offers a useful graph-based tool for exploring RNA motifs and suggesting large RNA motifs for design.

Original languageEnglish (US)
Pages (from-to)811-821
Number of pages11
JournalJournal of Molecular Biology
Volume428
Issue number5
DOIs
StatePublished - Feb 27 2016

Keywords

  • Prediction of RNA-like motifs
  • RNA atlas
  • RNA design
  • RNA motifs
  • RNA secondary structure

ASJC Scopus subject areas

  • Structural Biology
  • Molecular Biology

Fingerprint Dive into the research topics of 'Predicting Large RNA-Like Topologies by a Knowledge-Based Clustering Approach'. Together they form a unique fingerprint.

  • Cite this