Enhancing genomic disorder prediction through Feynman Concordance and Interpolated Nearest Centroid techniques

Area of Science:

• Genomics
• Biomedical Research
• Computational Biology

Background:

• Genomic technologies offer extensive clinical biomedical applications for enhancing healthcare.
• Genome disorder prediction is crucial for identifying multivariate diseases like cancer and diabetes.
• Current machine and deep learning methods for genome prediction lack sufficient accuracy.

Purpose of the Study:

• To propose a novel method, Quadratic Feynman Polynomial Interpolated and Vector Nearest Centroid-based (QFPI-VNC), for accurate genome disorder prediction.
• To improve sensitivity and specificity in forecasting genome disorders.
• To address the limitations of existing genome prediction techniques.

Main Methods:

• Utilized medical data from a public genomes dataset for children.
• Applied Linear Quadratic and Feynman Kac Genome filtering for efficient results.
• Employed Concordance Correlated Polynomial Interpolation for accurate genome-wide data extraction.
• Fused extracted features and used a Support Vector and Nearest Centroid model for prediction.

Main Results:

• The QFPI-VNC method demonstrated prospective performance compared to state-of-the-art methods.
• Achieved higher genome disease detection rate (14%), accuracy (11%), sensitivity (14%), and specificity (12%).
• Showcased a significant reduction in convergence speed (29%).

Conclusions:

• The proposed QFPI-VNC method offers improved sensitivity and specificity for genome disorder prediction.
• QFPI-VNC presents a promising advancement in early detection and management of genetic diseases.
• The method's enhanced performance suggests its potential for widespread clinical biomedical applications.