ATTIC is a novel stacked-ensemble learning model to identify A-to-I editing sites in three species accurately. We first comprehensively evaluated 37 RNA sequence-derived features combined with 14 popular machine learning algorithms. Then we selected the optimal base models to build a series of stacked ensemble models. The final ATTIC framework was developed upon the optimal models improved by feature selection strategy for specific species. Extensive cross-validation and independent tests demonstrate that ATTIC outperforms state-of-the-art tools for predicting A-to-I editing sites.
The environment on our computer is as follows:
- Python 3.8
- pycaret 2.3.5
- Pandas 1.4.1
- NumPy 1.21.2
- scikit-learn 0.23.2
python ATTIC.py -i "input fasta file" -s "species" ['H.sapiens','M.musculus','D.melanogaster'] -o "output file" -t "threshold"For example:
python ATTIC.py -i "Example/input.fasta" -s 'H.sapiens' -o Example -t 0.5The predicted sequence length of H. sapiens and D. melanogaster must be greater than 51, while the expected sequence length of M. musculus must be greater than 41.
Please cite our paper if you use this code in your own work.
Feel free to contact us if you have any questions. You can send your queries and suggestions to the developers: fuyi.li@nwafu.edu.cn , Jiangning.song@monash.edu and lachlan.coin@unimelb.edu.au .