IPPF-FE：An integrated peptide and protein function prediction framework based on fused features and ensemble models

We are designing user-friendly website for biologists!

1. Introduction

This repository contains source data and code for paper "An integrated peptide and protein function prediction framework based on fused features and ensemble models". IPPF-FE is a python implementation of the model.

2. Installation

python=3.6.9

You could configure enviroment by running this:

pip install -r requirment.txt

Notice:

1. You need install pretrained language modoel ProtT5-XL-UniRef50, the link is provided on ProtT5-XL-U50.
2. You also could pip install torch 1.10.1+cu113 by manual method, the link is provided as on Pytorch.

3. Requirments

In order to run successfully, the embedding of ProtT5-XL-UniRef50 requires GPU. We utilized an NVIDIA GeForce RTX 3080 with 10018MiB to embed peptide or protein sequences to 1024-dimensional vector. And Hand-crafted features could be implemented on personal computer. Other hardware equipments are not necessary.

4. Usage

For each dataset, you could run corresponding .py file, train model and external test are all implemented. We took Antibacterial peptides dataset as an example.

Train model

python Pantibacterial.py Train -

External test

python Pantibacterial.py Test test.fasta

Optional input variables

-predicted type    e.g. Pantibacterial.py, Phemolytic.py, Pbiofilm_inhibitory.py, PDPP_IV.py, PT3SEs.py, Pcsq_resolution.py, Pcsq_rfree.py, Pgpl.py, Pcyclinp.py
-Train or Test    e.g. Train, Test
-input file    e.g. test.fasta

5. References

Please cite the paper IPPF-FE：An integrated peptide and protein function prediction framework based on fused features and ensemble models.

6. Contact

If you have any question, you could contact han.yu@siat.ac.cn.