Poster Presentation 50 Years Shine-Dalgarno Symposium 2023

PRIMITI: A machine learning model for identifying novel miRNA-target mRNA interactions (#142)

Korawich Uthayopas 1 2 3 , Alex G. C. de Sá 1 2 3 4 , Azadeh Alavi 5 , Douglas E. V. Pires 1 2 3 6 , David B. Ascher 1 2 3 4
  1. Systems and Computational Biology, Bio21 Institute, University of Melbourne, Parkville, Victoria, 3052, Australia
  2. School of Chemistry and Molecular Biosciences, University of Queensland, Brisbane , Queensland, 4072 , Australia
  3. Computational Biology and Clinical Informatics, Baker Heart and Diabetes Institute, Melbourne, Victoria, 3004, Australia
  4. Baker Department of Cardiometabolic Health, Melbourne Medical School, University of Melbourne, Parkville, Victoria, 3010, Australia
  5. School of Computational Technology, RMIT University, Melbourne, Victoria, 3000, Australia
  6. School of Computing and Information Systems, University of Melbourne, Parkville , Victoria, 3052, Australia

The vast majority of human genes are suppressed post-transcriptionally by miRNAs, a small non-coding RNA that plays a crucial role in human gene regulation1. Multiple studies have demonstrated a strong correlation between miRNA dysregulation and a broad spectrum of diseases, including cancers, neurological disorders, and cardiovascular diseases2-3. A comprehensive map of miRNA functions, miRNA-target repression, is necessary to advance our knowledge of miRNA biological function and their clinical applications4-5.

Computational models have been developed to facilitate the identification of miRNA-target mRNA suppression activities6-8. However, the existing models have exhibited limited predictive capability and applicability. In this study, we developed PRIMITI, a PRedictive model for identifying novel MIRNA-Target mRNA Interactions. Machine learning was implemented to predict miRNA-mRNA suppression activities by modelling molecular interactions between miRNAs and target sites in 3’-untranslated regions (3’-UTR). We introduced a negative sample selection to mitigate a publication bias, improving the reliability of negative samples. PRIMITI additionally combines information on single polymorphisms (SNPs) and physicochemical properties to improve target site characterisation.

The model achieved robust predictive performance, with Matthews correlation coefficients of up to 0.81 for miRNA-target site identification and up to 0.77 for mRNA repression activity prediction on cross-validation and independent blind test sets. PRIMITI was evaluated further using an external blind microarray dataset9 and experimentally validated lists of miRNA-mRNA interactions10-11, outperforming existing state-of-the-art approaches. PRIMITI is publicly available through a user-friendly web server at https://biosig.lab.uq.edu.au/primiti, making it an invaluable resource for preliminary screening of miRNA-mediated repression.

  1. O'Brien, J., et al., Overview of microrna biogenesis, mechanisms of actions, and circulation. Frontiers in Endocrinology, 2019. 9.
  2. Hill, M. and N. Tran, miRNA interplay: mechanisms and consequences in cancer. Dis Model Mech, 2021. 14(4).
  3. Qiu, L., Tan, E.K., and Zeng, L., MicroRNAs and Neurodegenerative Diseases. Advances in Experimental Medicine and Biology, 2015. p. 85-105.
  4. Hanna, J., Hossain, G.S., and Kocerha, J., The Potential for microRNA Therapeutics and Clinical Research. Front Genet, 2019. 10: p. 478.
  5. Gebert, L.F., et al., Miravirsen (SPC3649) can inhibit the biogenesis of miR-122. Nucleic Acids Res, 2014. 42(1): p. 609-21.
  6. Paraskevopoulou, M.D., et al., DIANA-microT web server v5.0: service integration into miRNA functional analysis workflows. Nucleic Acids Res, 2013. 41(Web Server issue): p. W169-73.
  7. Agarwal, V., et al., Predicting effective microRNA target sites in mammalian mRNAs. Elife, 2015. 4.
  8. Wong, N. and X. Wang, miRDB: an online resource for microRNA target prediction and functional annotations. Nucleic Acids Res, 2015. 43(Database issue): p. D146-52.
  9. Linsley, P.S., et al., Transcripts targeted by the microRNA-16 family cooperatively regulate cell cycle progression. Mol Cell Biol, 2007. 27(6): p. 2240-52.
  10. Huang, H.Y., et al., miRTarBase 2020: updates to the experimentally validated microRNA-target interaction database. Nucleic Acids Res, 2020. 48(D1): p. D148-D154.
  11. Karagkouni, D., et al., DIANA-TarBase v8: a decade-long collection of experimentally supported miRNA-gene interactions. Nucleic Acids Res, 2018. 46(D1): p. D239-D245.