eleventh83 / deepMiRGene

DeepMiRGene

deepMiRGene

Deep Recurrent Neural Network-Based Identification of Precursor microRNAs

Seunghyun Park, Seonwoo Min, Hyun-Soo Choi, and Sungroh Yoon, in Proceedings of the Annual Conference on Neural Information Processing Systems (NIPS), Long Beach, USA, December 2017.

http://papers.nips.cc/paper/6882-deep-recurrent-neural-network-based-identification-of-precursor-micrornas

Usage

python inference/deepMiRGene.py -i <input file> -o <output file>

• input file: fasta format
• output file: 0 (true pre-miRNA) or 1 (pseudo pre-miRNA)

Dependencies

1. biopython

2. RNAfold (python version)

3. sklearn

4. Keras

• theano backended

Reproduce

1. reproduce/cv.py

• cross-validation results for the human and cross-species dataset (Table 2)

2. reproduce/test.py (human and cross-species)

• test results for the human and cross-species dataset (Table 2)

3. reproduce/test_new.py (new)

• test results for the new dataset (Table 3)