Machine learning based predictive model and systems-level network of host-microbe interactions in post-COVID-19...

Area of Science:

• Medical Mycology
• Infectious Diseases
• Computational Biology

Background:

• Mucormycosis, a rare fungal infection, is increasingly associated with Coronavirus disease (COVID-19), particularly in India.
• Cases involve rhino-cerebral-orbital, angioinvasive, pulmonary, and gastrointestinal manifestations.
• Factors like immune dysregulation, steroid use, IL-6 therapies, and mechanical ventilation in COVID-19 patients contribute to mucormycosis development.

Purpose of the Study:

• To analyze the presentation, course, diagnosis, and treatment of post-COVID-19 mucormycosis.
• To identify risk factors, drug targets, and biomarkers for this opportunistic infection.
• To develop predictive models for post-COVID-19 mucormycosis using machine learning and bioinformatics.

Main Methods:

• Machine learning, systems biology, and bioinformatics analyses were employed.
• Text mining identified key characteristics of post-COVID-19 mucormycosis cases.
• A host-microbe interaction network was constructed to identify candidate molecules.

Main Results:

• Candidate characteristics of post-COVID-19 mucormycosis were identified, including demographics, symptoms, and clinical parameters.
• A machine learning model demonstrated predictive potential for these characteristics.
• Replicase polyprotein 1a (R1A) was identified as a potential drug target, and Ribosomal Protein S6 (RPS6) as a potential biomarker.

Conclusions:

• Identified risk factors, drug targets, and biomarkers can aid in the prognosis, early diagnosis, and therapy of post-COVID-19 mucormycosis.
• Further experimental validation is required for these findings.
• This research contributes to understanding and managing opportunistic infections in COVID-19 survivors.