Genetic justification of COVID-19 patient outcomes using DERGA, a novel data ensemble refinement greedy algorithm

Area of Science:

• Immunology
• Genetics
• Computational Biology

Background:

• Complement inhibition shows therapeutic potential in diseases like COVID-19.
• Accurate prediction of disease severity is essential for patient management.
• Identifying genetic factors influencing complement activity is critical for personalized medicine.

Purpose of the Study:

• To identify key complement genetic variants associated with COVID-19 outcomes.
• To develop an AI-based prediction tool for disease severity.
• To determine an optimal pattern of genetic variants for accurate outcome prediction.

Main Methods:

• Analysis of genetic data from 204 hospitalized COVID-19 patients.
• Utilized an artificial intelligence (AI)-based algorithm for outcome prediction (ICU vs. non-ICU admission).
• Employed the alpha-index to identify predictive genetic variants and the DERGA algorithm for pattern determination.

Main Results:

• Identified 30 highly predictive complement genetic variants using the alpha-index.
• The DERGA algorithm achieved 97% accuracy in predicting COVID-19 disease outcome.
• Detected a total of 977 variants, with individual variations ranging from 40 to 161 per patient.

Conclusions:

• The alpha-index is effective for ranking numerous genetic variants.
• AI-driven analysis of complement genetic variants enables highly accurate prediction of disease outcomes.
• This approach facilitates the development of predictive tools for personalized COVID-19 management.