Differential gene expression profiling reveals potential biomarkers and pharmacological compounds against SARS-CoV-2: insights from machine learning and bioinformatics approaches

Abstract

The precise role of severe acute respiratory syndrome coronavirus 2 in the pathophysiology of the nasopharyngeal tract (NT) is still unfathomable. Therefore, we used the machine learning methods to analyze 22 RNAseq datasets from COVID 19 patients (n=8), recovered individuals (n=7), and healthy individuals (n=7) to find disease-related differentially expressed genes (DEGs). In comparison to healthy controls, we found 1960 and 153 DEG signatures in COVID 19 patients and recovered individuals, respectively. We compared dysregulated DEGs to detect critical pathways and gene ontology (GO) connected to COVID 19 comorbidities. In COVID-19 patients, the DEG miRNA and DEG transcription factors (TFs) interactions network analysis revealed that E2F1, MAX, EGR1, YY1, and SRF were the most highly expressed TFs, whereas hsa-miR-19b, hsa-miR-495, hsa miR 340, hsa miR 101, and hsa-miR-19a were the overexpressed miRNAs. Three chemical agents (Valproic Acid, Alfatoxin B1, and Cyclosporine) were abundant in COVID 19 patients and recovered individuals. Mental retardation, mental deficit, intellectual disability, muscle hypotonia, micrognathism, and cleft palate were the significant diseases associated with COVID-19 by sharing DEGs. Finally, we detected DEGs impacted by severe acute respiratory syndrome coronavirus 2 infection and mediated by TFs and miRNA expression, indicating that severe acute respiratory syndrome coronavirus 2 infection may contribute to various comorbidities. These pathogenetic findings can provide some crucial insights into the complex interplay between COVID 19 and the recovery stage and support its importance in the therapeutic development strategy to combat against COVID 19 pandemic.

Explainable artificial intelligence model for identifying COVID-19 gene biomarkers

“…In current research, we found that an increase in the level of IFI27 for COVID-19 increases the risk of disease. In the literature, they reported that SARS-CoV-2 infection suppressed the expression of the DCUN1D3 gene and the disease was associated with this gene [ 48 ]. Our research found that low levels of the DCUN1D3 gene are associated with COVID-19.…”

A systematic review of artificial intelligence-based COVID-19 modeling on multimodal genetic information

“…Moreover, nsp3 was more conserved among the SARS family than 15 other variants of coronavirus ( Ong et al, 2020 ). Table 1 contains the clinical biomarkers, methods, and objectives gathered from AI based COVID-19 literature ( Downes et al, 2021 ; Saberi-Movahed et al, 2022 ; Fujisawa et al, 2021 ; Liu et al, 2021 ; Zeng et al, 2020 ; Ke et al, 2020 ; Joshi et al, 2021 ; Sardar et al, 2021 ; Hoque et al, 2022 ).…”

“… Ref. Methods Objectives Biomarkers Downes et al (2021) SpliceAI Risk loci identification LZTFL1 Saberi-Movahed et al (2022) Matrix Factorization Prognostic biomarker ABG, CRP Fujisawa et al (2021) PCA Gene Associations NF-kappaB, H3k36me3 Liu et al (2021) PCA Biomarker Discovery UCHL1 Zeng et al (2020) CoV-KGE Model Drug repurposing 41 repurposable drugs Ke et al (2020) Deep Neural Network Drug repurposing 13 potential drug candidates Joshi et al (2021) RNN Drug Development Palmatine and Sauchinone stability Sardar et al (2021) Random Forest Potential Drug Targets VGFR2, EGFR, FA7, ANGP2 Hoque et al (2022) BORUTA Transcriptomic Signatures PSMB8, COLCA2, FAM83A, LGALS3BP, IRF9 …”