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

Mucormycosis, a rare infection is caused by fungi Mucorales. The affiliation of mucormycosis with Coronavirus disease (COVID-19) is a rising issue of concern in India. There have been numerous case reports of association of rhino-cerebral-orbital, angioinvasive, pulmonary, respiratory and gastrointestinal tract related mucormycosis in patients with history of COVID-19. The immune dysregulation, preposterous use of steroids, interleukin-6-directed therapies and mechanical ventilation in COVID-19 immunocompromised individuals hypothesizes and predisposes to advancement of mucormycosis. The gaps in mode of presentation, disease course, diagnosis and treatment of post-COVID-19 mucormycosis requires critical analysis in order to control its morbidity and incidence and for prevention and management of opportunistic infections in COVID-19 patients. Our study performs machine learning, systems biology and bioinformatics analysis of post-COVID-19 mucormycosis in India incorporating multitudinous techniques. Text mining identifies candidate characteristics of post-COVID-19 mucormycosis cases including city, gender, age, symptoms, clinical parameters, microorganisms and treatment. The characteristics are incorporated in a machine learning based disease model resulting in predictive potentiality of characteristics of post-COVID-19 mucormycosis. The characteristics are used to create a host-microbe interaction disease network comprising of interactions between microorganism, host-microbe proteins, non-specific markers, symptoms and drugs resulting in candidate molecules. R1A (Replicase polyprotein 1a) and RPS6 (Ribosomal Protein S6) are yielded as potential drug target and biomarker respectively via potentiality analysis and expression in patients. The potential risk factors, drug target and biomarker can serve as prognostic, early diagnostic and therapeutic molecules in post-COVID-19 mucormycosis requiring further experimental validation and analysis on post-COVID-19 mucormycosis cases.

In-silico study of peptide-protein interaction of antimicrobial peptides potentially targeting SARS and SARS-CoV-2 nucleocapsid protein.

This study is an attempt to find a suitable therapy using antimicrobial peptides (AMPs) by identifying peptide-protein interaction of AMPs and nucleocapsid protein of SARS and SARS-CoV- 2. The AMPs were shortlisted from the APD3 database (Antimicrobial peptide database) based on various physicochemical parameters. The binding efficacy of AMPs was measured using the lowest energy score of the docked complexes with 10 selected AMPs. For SARS-CoV, AP00180 showed the best pose with a binding affinity value of - 6.4 kcal/mol. Prominent hydrogen bonding interactions were observed between Lys85 (nucleocapsid receptor) and Arg13 (antimicrobial peptide ligand) having the least intermolecular distance of 1.759 Å. For SARS-CoV-2, AP00549 was docked with a binding affinity value of - 3.4 kcal/mol and Arg119 and Glu14 of receptor nucleocapsid protein and ligand AMP having the least intermolecular distance of 2.104 The dynamic simulation was performed at 50 ns to check the stability of the final docked complexes, one with each protein. The two best AMPs were AP00180 (Human Defensin-5) for SARS and AP00549 (Plectasin) for SARS-CoV-2. From positive results of dynamic simulation and previously known knowledge that some AMPs interact with the nucleocapsid of coronaviruses, these AMPs might be used as a potential therapeutic agent for the treatment regime of SARS-CoV-2 and SARS infection. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s40203-021-00103-z.