trnamod

trnamod predicts post-transcriptional uridine modifications in transfer RNA (tRNA), specifically pseudouridine (Y) and dihydrouridine (D), using SVM-based hybrid models trained on MODOMICS datasets to provide kingdom-wise modification predictions.


Key Features:

  • Prediction of Uridine Modifications: Predicts pseudouridine (Y) and dihydrouridine (D) occurrences in tRNA sequences.
  • Kingdom-Wise Prediction Models: Implements a kingdom-wise approach with a three-step strategy that incorporates common and individualized datasets for domain-specific models.
  • Hybrid Modeling: Combines binary and structural information in a hybrid feature representation to improve predictive accuracy.
  • SVM-Based Classification and Performance: Trains Support Vector Machine (SVM) classifiers reporting Area Under the Curve (AUC) values of 0.936–0.987.
  • Dataset and Validation: Constructs a common model using MODOMICS-2008 with five-fold cross-validation and performs subsequent evaluation and individual model development using MODOMICS-2012.
  • Input Data Types: Supports predictions from both tRNA sequences and whole genome data.
  • Modification Classification: Classifies predicted uridine modifications into specific modification types.

Scientific Applications:

  • Genetic Regulation and Expression: Facilitates studies of how tRNA uridine modifications affect genetic regulation and gene expression.
  • Genome Architecture and Protein Synthesis: Enables investigation of the implications of uridine modifications on genome architecture, codon usage, and protein synthesis.
  • Comparative Genomics and Evolutionary Biology: Supports comparative and evolutionary analyses of tRNA modification patterns across biological kingdoms.

Methodology:

Uses a three-step strategy with a common model built from MODOMICS-2008 validated by five-fold cross-validation, followed by performance evaluation and development of individual kingdom-wise models using MODOMICS-2012; employs hybrid binary and structural features and Support Vector Machine (SVM) classifiers with reported AUCs of 0.936–0.987.

Topics

Details

Tool Type:
web application
Operating Systems:
Mac, Linux, Windows
Added:
10/11/2022
Last Updated:
10/11/2022

Operations

Data Inputs & Outputs

Publications

Panwar B, Raghava GP. Prediction of uridine modifications in tRNA sequences. BMC Bioinformatics. 2014;15(1). doi:10.1186/1471-2105-15-326. PMID:25272949. PMCID:PMC4287530.

Documentation

Links