Pandurang Kolekar
Ph.D. Bioinformatics
Sr. Staff Bioinformatics Research Scientist
St. Jude Children's Research Hospital, Memphis, TN, USA
https://www.stjude.org/
http://orcid.org/0000-0003-0044-0076
Machine Learning
Prediction of Cellular and Viral Internal Ribosome Entry Site (IRES) using Support Vector Machine (SVM)
​
Cellular mRNAs are predominantly translated in a cap-dependent manner. However, some viral and a subset of cellular mRNAs initiate their translation in a cap-independent manner. This requires presence of a structured RNA element, known as, Internal Ribosome Entry Site (IRES) in their 5' untranslated regions (UTRs). Experimental demonstration of IRES in UTR remains a challenging task. Computational prediction of IRES merely based on sequence and structure conservation is also difficult, particularly for cellular IRES.
​
A web server, IRESPred is developed for prediction of both viral and cellular IRES using Support Vector Machine (SVM). The predictive model was built using 35 features that are based on sequence and structural properties of UTRs and the probabilities of interactions between UTR and small subunit ribosomal proteins (SSRPs). The model was found to have 75.51% accuracy, 75.75% sensitivity, 75.25% specificity, 75.75% precision and Matthews Correlation Coefficient (MCC) of 0.51 in blind testing. IRESPred was found to perform better than the only available viral IRES prediction server, VIPS.
Team members:
Dr. Pandurang Kolekar
Mr. Abhijeet Pataskar
Dr. Urmila Kulkarni-Kale
Dr. Jayanta Pal
Dr. Abhijeet Kulkarni