kaihe-plus / body-fluid-models

Prediction Models for Body-fluid Protein Discovery

Body-fluid proteins are very important because they are promising as biomarkers. This repository contains several prediction models for body-fluid protein discovery.

Models

Name	Description	Type	Script
SVM	Support Vector Machine	Supervised	svm.py
DT	Decision Tree	Supervised	dt.py
DNN	Deep Neural Network	Supervised	dnn.py
DeepSec	Deep Learning	Supervised	deepsec.py
MultiSec	Deep Learning + Multi-task learning	Supervised	multisec.py
PU-DT	DT + PU Bagging	Semi-supervised	pu-dt.py
PU-SVM	SVM + PU Bagging	Semi-supervised	pu-svm.py
MPUSec	Multi-task learning + PU Learning	Semi-supervised	mpusec.py

Data

Data can be downloaded from this link.

RUN

python script_name.py --folder data_dir

Reference

• Dan Shao, Lan Huang, et al. Computational Prediction of Human Body-Fluid Protein. BIBM 2019.
• Dan Shao, Lan Huang, et al. DeepSec: A Deep Learning Framework for Secreted Protein Discovery in Human Body Fluids. Bioinformatics 2021.
• Kai He, Yan Wang, et al. MultiSec: Multi-Task Deep Learning Improves Secreted Protein Discovery in Human Body Fluids. Mathematics 2022.
• Kai He, Yan Wang, et al. A Multi-task Positive-unlabeled Learning Framework to Predict Secreted Proteins in Human Body Fluids. Complex & Intelligent Systems 2023.