genbio-ai/rna-downstream-tasks

AIDO.RNA Benchmark Datasets

mRNA related tasks

• Translation efficiency prediction from Chu et al.(2024) [1]
  • 3 cell lines: Muscle, pc3, HEK
  • input sequence: 5'UTR
  • 10-fold cross-validation split
• mRNA expression level prediction from Chu et al.(2024) [1]
  • 3 cell lines: Muscle, pc3, HEK
  • input sequence: 5'UTR
  • 10-fold cross-validation split
• Mean ribosome load prediction from Sample et al. (2019) [2]
  • input sequence: 5'UTR
  • ouput: mean ribosome load
  • the original data source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE114002
  • Similar to the previous studies [2, 4], we also split the data into the following three
    • train: total 76.3k samples
    • val: total 7600 samples (also called as Random 7600 in [4])
    • test: total 7600 samples (also called as Human 7600 in [4])
• Transcript abundance prediction from Outeiral and Deane (2024) [3]
  • 7 organisms: A. thaliana, D. melanogaster, E.coli, H. sapiens, S. cerevisiae, H. volcanii, and P. pastoris
  • input sequence: CDS
  • 5-fold cross-validation split
• Protein abundance prediction from Outeiral and Deane (2024) [3]
  • 5 organisms: A. thaliana, D. melanogaster, E.coli, H. sapiens, and S. cerevisiae
  • input sequence: CDS
  • 5-fold cross-validation split
  • Note: We have transformed the label to logarithm space using the following function: log(1+x).

RNA function prediction tasks

The datasets listed below are collected following the setting in Wang et al. (2023) [4].

• Cross-species splice site prediction
  • 2 datasets: acceptor, donor
  • 4 test species: zebrafish, fruit fly, worm, and plant
  • input sequence: pre-mRNA fragment
• ncRNA family classification
  • 2 datasets: boundary noise 0, boundary noise 200
  • input sequence: small noncoding RNA with different level of boundary noise
• RNA modification site prediction
  • 12 labels (modification sites): Am, Cm, Gm, Tm, m1A, m5C, m5U, m6A, m6Am, m7G, Φ, and I.

Reference

1. Yanyi Chu, Dan Yu, Yupeng Li, Kaixuan Huang, Yue Shen, Le Cong, Jason Zhang, and Mengdi Wang. A 5 utr language model for decoding untranslated regions of mrna and function predictions. Nature Machine Intelligence, pages 1–12, 2024.
2. Paul J Sample, Ban Wang, David W Reid, Vlad Presnyak, Iain J McFadyen, David R Morris, and Georg Seelig. Human 5 utr design and variant effect prediction from a massively parallel translation assay. Nature biotechnology, 37(7):803–809, 2019.
3. Carlos Outeiral and Charlotte M Deane. Codon language embeddings provide strong signals for use in protein engineering. Nature Machine Intelligence, 6(2):170–179, 2024.
4. Xi Wang, Ruichu Gu, Zhiyuan Chen, Yongge Li, Xiaohong Ji, Guolin Ke, and HanWen. Uni-rna: universal pre-trained models revolutionize rna research. bioRxiv, pages 2023–07, 2023.