Piwi-interacting RNAs (piRNAs) Prediction Using Deep Learning
- 0 Collaborators
Piwi-interacting RNAs (piRNAs) are small sequences of non-coding RNAs that have several functions related to the protection and regulation of genes and genomes. However, due to the lack of conservation of the sequences and their structural elements, their identification is a complicated task. This project aims at using neural network models, such as deep learning and generative models, to extract information from the massive amount of data related to piRNA sequences, in order to develop a robust classification methods for this type of non-coding RNA sequences. ...learn more
Project status: Under Development
Intel Technologies
Intel Opt ML/DL Framework
Overview / Usage
Piwi-interacting RNAs (piRNAs) are small sequences of non-coding RNAs that have several functions related to the protection and regulation of genes and genomes. However, due to the lack of conservation of the sequences and their structural elements, their identification is a complicated task. This project aims at using neural network models, such as deep learning and generative models, to extract information from the massive amount of data related to piRNA sequences, in order to develop a robust classification methods for this type of non-coding RNA sequences.
Methodology / Approach
We use Bioinformatic tools to extract features from a massive amount of non-coding RNA sequences, classifying them into piRNAs and non-piRNAs. With such amount of data, we investigated deep learning models using dropout and advanced learning algorithms, and also models using stacked auto-encoders and restricted boltzmann machines, in order to construct robust classifiers for these sequences.
Technologies Used
Python 3.6.5
gcc 4.8.5
keras 2.1.5
theano 1.0.1
numpy 1.12.1
scikit-learn 0.19.1
scipy 1.0.1
Memory RAM capacity: 15.51361 GB
Processor model name: Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz