| Literature DB >> 28641399 |
Shahidul Islam1, Ping Ge1, Shaojie Zhang1.
Abstract
The analysis of RNA tertiary structure is hindered by the fact that not too many structural data are available and a significant amount of them are in low resolution. Due to the atomic coordinate errors posed by the limitations of low-resolution RNA three-dimensional structures, it becomes a critical challenge to extract key geometric characteristics of RNA, particularly, the interaction of bases. To address this issue, we have devised a comparative method, named CompAnnotate, that utilizes more precise structural information of high-resolution homologs to annotate the base-pairing interactions in the low-resolution structures, by aligning and making comparative geometric assessments. The benchmarking results show that our method can improve the annotations of the existing methods significantly. We have achieved different levels of improvements for various methods and datasets, including an example of significant sensitivity and precision enhancement from 28 to 57% and from 53 to 82%, respectively.Entities:
Mesh:
Substances:
Year: 2017 PMID: 28641399 PMCID: PMC5737500 DOI: 10.1093/nar/gkx538
Source DB: PubMed Journal: Nucleic Acids Res ISSN: 0305-1048 Impact factor: 16.971