CORAIN is constructed to encode RNA of (1) providing various coding strategies for RNA from three perspectives, sequence, structure and physical chemical properties, (2) offering the combined coding strategies of RNA-RNA interaction, RNA-protein interaction and RNA-drug interaction, (3) covering the encoding methods not only for conventional machine learning but also the deep learning. Therefore, users can get the encoding results of RNA or the encoding results of RNA interaction pairs from CORAIN, which can be directly used for artificial intelligence.

Contents

• Contents
• Workflow
• Requirements and installment
• Basic usage
• Study demo
  • 1. RNA-only
  • 2. RNA-RNA interaction
  • 3. RNA-protein interaction
  • 4. RNA-compound interaction
• Question && issue

Workflow

CORAIN coding and combined feature evaluation are divided into two parts. Step 1: Encode the uploaded RNA and RNA-interacting molecules (RNA, protein, small molecule compounds) to obtain one-dimensional or two-dimensional coding features; Step 2: If one-dimensional coding features are performed, and the sample type If it is greater than or equal to 2, then the combined feature evaluation process can be performed.

Requirements and installment

This software is developed with Python 3.X, Python 3.X is required as runtime environment. we recommemd user to install anaconda3 which is well-known as a scientific Python.

git clone https://github.com/idrblab/corain
cd corain
# virtual environment are recommended
conda create -n corain python=3.8
pip install -r requirements.txt

Basic usage

Users can use corain to (1) encode RNA and (2) evaluate features.

python corain.py --help

usage: corain.py [-h] [-t TYPE] [-a AFASTAPATH] [-b BFASTAPATH] 
[-l INTERFILEPATH] [-o RESULTPATH] [-d DIMENSION] [-s SAVETYPE] 
[-n N_SELECT] [-c COM_NUM] [-m MODELNM]

optional arguments:
  -h, --help            show this help message and exit
  -t TYPE, --type TYPE  The encoding file type: RNAonly, RNA-RNA, RNA-pro, 
                        RNA-compound. It is the coding task of a single 
                        RNA fasta file, or RNA interaction task.    
  -a AFASTAPATH, --Afastapath AFASTAPATH
                        The RNA fasta A file path. RNA fasta format is necessary 
                        and standard fasta format can be seen
                        https://blast.ncbi.nlm.nih.gov/Blast.cgi.
  -b BFASTAPATH, --Bfastapath BFASTAPATH
                        The encoding fasta B file. If the type is RNAonly, 
                        this file is not necessary.
  -l INTERFILEPATH, --Interfilepath INTERFILEPATH
                        The label file. This file format is .csv format.
  -o RESULTPATH, --Resultpath RESULTPATH
                        The encoding result file path.
  -d DIMENSION, --dimension DIMENSION
                        The feature dimension of encoding feature: 1 or 2.
  -s SAVETYPE, --savetype SAVETYPE
                        The encoding result file type: csvnpy or csv or 
                        npy, defatult is csv and npy.
  -n N_SELECT, --n_select N_SELECT
                        The encoding feature number: 1-10. Default is all 
                        encoding features.
  -c COM_NUM, --com_num COM_NUM
                        Number of combination features: 1-10. Default is all 
                        encoding features combinations (1023).
  -m MODELNM, --modelnm MODELNM
                        Classification model of evaluation feature 
                        combination:'RF','svm','xgboost','DNN','CNN'.

The paramter "n_select" is corresponsed with parameter "dimension" as following table.

n_select	1D	2D
1	Open reading frame	Sequence-intrinsic Features:One-hot encoding
2	Entropy density of transcript	Sparse encoding
3	Global descriptor	Structure-based Features:One-hot encoding
4	K-mer	-
5	Guanine-cytosine/Codon related	-
6	Pse-protein/Nucleotide	-
7	Nucleotide related	-
8	EIIP based spectrum	-
9	Lipoaffinity index	-
10	Partition coefficient	-
11	Molar refractivity	-
12	Hydrogen bond related	-
13	Secondary structure	-
14	Topological indices	-
15	Molecular fingerprint	-
16	3D Structure-based	-

Study demo

1. RNA-only

Users can run RNA-only tasks like following shell script.

python corain.py -t RNAonly \
    -a ./demo/RNA-only/RNA.fasta \
    -l ./demo/RNA-only/RNA_label.csv \
    -o ./out \
    -d 1 \
    -s csvnpy \
    -n 4 \
    -c 8 \
    -m svm

2. RNA-RNA interaction

python corain.py -t RNA-RNA \
    -a ./demo/RNA-RNA/SampleData-lncRNA-A.fasta \
    -b ./demo/RNA-RNA/SampleData-miRNA-B.fasta \
    -l ./demo/RNA-RNA/RNA-RNA-Interacting.csv \
    -o ./out

3. RNA-protein interaction

python corain.py -t RNA-pro \
    -a ./demo/RNA-pro/SampleData-RNA-A.fasta \
    -b ./demo/RNA-pro/SampleData-Protein-B.fasta \
    -l ./demo/RNA-pro/RNA-Protein-Interacting.csv \
    -o ./out

4. RNA-compound interaction

python corain.py -t RNA-compound \
     -a ./demo/RNA-compound/SampleData-RNA-A.fasta \
     -b ./demo/RNA-compound/Small-Moleculer.smi \
     -l ./demo/RNA-compound /RNA-small-molecule-Interacting.csv \
     -o ./out

Question && issue

If there are any usage problems, welcome to submit issues to this repository or contact author with email [email protected]