OmicPredict: a framework for omics data prediction using ANOVA-Firefly algorithm for feature selection.

Area of Science:

• Computational biology
• Bioinformatics
• Machine learning in healthcare

Background:

• High-throughput omics data analysis is crucial for understanding complex diseases.
• Machine learning models are increasingly used for early disease prediction.
• Developing robust predictive models for multiple diseases from omics data remains a challenge.

Purpose of the Study:

• To propose and validate a novel framework, 'OmicPredict', for predicting multiple diseases using omics data.
• To integrate a hybrid feature selection method with a deep neural network (DNN) for enhanced predictive accuracy.
• To demonstrate the framework's efficacy across diverse diseases including Alzheimer's, breast cancer, and COVID-19.

Main Methods:

• Developed a hybrid feature selection method combining Analysis of Variance (ANOVA) and the firefly algorithm.
• Employed a deep neural network (DNN) model within the OmicPredict framework.
• Applied the framework to three distinct omics datasets: GSE33000/GSE44770 (Alzheimer's), METABRIC (Breast Cancer HER2+), and GSE157103 (COVID-19).

Main Results:

• The DNN model achieved a high Area Under Curve (AUC) of 0.949 for Alzheimer's disease prediction.
• The framework demonstrated excellent performance for breast cancer (AUC=0.987) and COVID-19 (AUC=0.989).
• OmicPredict outperformed established models like Random Forest and Naïve Bayes, as well as existing research benchmarks.

Conclusions:

• The OmicPredict framework provides an effective approach for multi-disease prediction using omics data.
• The hybrid feature selection and DNN model combination significantly improves predictive accuracy.
• This work contributes to advancing early disease detection and personalized treatment strategies.