Predicting ncRNA-protein interaction using high-level features

we proposed a computational method IPMiner to predict ncRNA-protein interactions from sequences, which made use of deep learning and further improve its performance using stacked ensembling. It automatically extracted high-level features from conjoint triad features of protein and RNA sequence using stacked autoencoder, then the high-level features are fed into random forest to predict ncRNA-protein interaction. Finally stacked ensembling is used to integrate different predictors to further improve prediction performance.

Dependency:
python 2.7
deep learning lib keras: https://github.com/fchollet/keras/ (version Keras-0.1.2 and the backend theano v0.9)
machine learning lib scikit-learn v0.17: https://github.com/scikit-learn/scikit-learn

Usage: python IPMiner.py -datatype=RPI488
where RPI488 is lncRNA-protein interaction dataset, and IPMiner will do 5-fold cross-validation for it. you can also choose other datasets, such as RPI1807, RPI369,, RPI13254 and NPInter.

python IPMiner.py -r=RNA_fasta_file -p=protein_fasta_file
it will predict pairwise interaction score for RNAs and protiens in input fasta file.

Reference
Xiaoyong Pan, Yong-Xian Fan, Junchi Yan and Hong-Bin Shen. IPMiner: hidden ncRNA-protein interaction sequential pattern mining with stacked autoencoder for accurate computational prediction. BMC Genomics. 2016, 17:582 DOI: 10.1186/s12864-016-2931-8.

Contact: xypan172436atgmail.com