midisec / PINC

PINC (Plant Non-Coding Recognition Tool) is a powerful tool for identifying non-coding RNAs by analyzing k-mer frequency, cds, sequence length and GC content through sequence intrinsic composition to effectively differentiate between protein-coding and non-coding RNAs for a growing number of non-model plants.

PINC

A powerful tool for identifying non-coding RNAs in plants by analysing k-mer frequency, cds-related features, sequence length and GC content to distinguish between the growing number of non-coding RNAs and coding RNAs in plants.

Features

• High precision (ensemble learning)
• Multiple high-performance base models
• Convenience of use
• Automated Forecasting
• Web Online

Documents

Documentation

Get Start

There are multiple ways to run this tool, feel free to choose one of the following method.

Run PINC from Web Online (Fastest)

http://www.pncrna.com/

Run PINC from docker (Locally、Simply)

1. Download the PINC and Add the data file to the project directory.

git clone https://github.com/midisec/PINC
cd PINC
# upload the data file (example: data.fasta)

All input data must be in fasta format

2. Pull and build the environment image. (Time required)

sudo docker build -t pinc_images .

3. Create and Enter a new container.

sudo docker run -it pinc_images bash

4. Execute PINC for prediction

python pinc.py -f data.fasta

Run PINC from source code (Complex)

1. Installation Environment(Autogluon、kentUtils)

2. Clone project, install related dependencies

git clone https://github.com/midisec/PINC
cd PINC
pip3 install -r requirements.txt

3. Execute PINC for prediction

python pinc.py -f data.fasta

Usage

Command line version

Prediction

python pinc.py -f <data.fasta>

Website online version

Prediction

After this, you will get a task page address with the uuid.

After that you can also check the history of the task by the uuid, usually it will be saved for one month.

View Results and Download results

Introduction

The Algorithm Framework/Process

The DataSets

The Training set data and validation set data (7 : 3).

Species	Coding	Non-coding	Total
Arabidopsis thaliana	2000	2000	4000
Glycine max	2000	2000	4000
Oryza sativa	2000	2000	4000
Vitis vinifera	2000	2000	4000
Total	8000	8000	16000

The Testing set data.

Species	Coding	Non-coding	Total
Cicer arietinum	2099	2099	4198
Gossypium darwinii	5622	5622	11244
Lactuca sativa	4682	4682	9364
Manihot esculenta	2808	2808	5616
Musa acuminata	2059	2063	4122
Nymphaea colorata	1708	1708	3416
Solanum tuberosum	8282	8282	16564
Sorghum bicolor	8657	8657	17314
Zea mays	7406	7406	14812
Total	51323	51327	102650

In the test set, the accuracy of the PINC ranged from 92.74% to 96.42%.

Citations

@article{zhang2022pinc,
  title={PINC: A Tool for Non-Coding RNA Identification in Plants Based on an Automated Machine Learning Framework},
  author={Zhang, Xiaodan and Zhou, Xiaohu and Wan, Midi and Xuan, Jinxiang and Jin, Xiu and Li, Shaowen},
  journal={International Journal of Molecular Sciences},
  volume={23},
  number={19},
  pages={11825},
  year={2022},
  publisher={MDPI}
}

Contributors