Network Based NLP Approach to Reveal Disease Comorbidities. COVID Case Study

Taylan , Emre
Pir, Pinar
SARS-CoV-2 virus, which caused the pandemic COVID-19 , infects the host cells via ACE2 receptor protein binding to spike proteins. Viral binding also alters the vascular effects of the ACE/ACE2 ratio, which results in a number of complications. These complications may have detrimental consequences on patients with chronic diseases such as diabetes, heart disease, tension-related disorders, and kidney disease, which already have an unbalanced ratio of ACE/ACE2. Due to this mechanism, COVID-19 is extremely comorbid, and there may be other comorbidities with yet unidentified etiologies. In this work, possible comorbidities of COVID-19 examined by a network-based representation of miRNA-gene-disease interactions to reveal possibly similarly altered mechanisms in between diseases. The original algorithm was implemented as the mpDisNet package which is a word-embedding based deep learning technique. Algorithm itself was modified and optimized to increase True-Positive discovery rate by removal of highly frequent diseases such as cancer and expanding the test data by manual curation of comorbid diseases. Original approach was also modified by integration of miRNA expression profiles from SARS-CoV-2 infected cell line and transcription factors regulated by these miRNAs from related databases. The main aim this work is discovery of possible unknown comorbidities by connecting the diseases by their miRNA mediated regulatory interactions. The algorithm can predict the majority of COVID-19's known comorbidities reported in CDC website with high scores; (Diabetes (0,996), Heart Diseases (0,989), Schizophrenia (0,994), and Hypertension (0,994)), as well as total of 131 diseases that have yet to be discovered to be comorbid with COVID-19. These potential comorbid diseases should be investigated further in order to raise COVID-19 awareness and prevention. Our approach can be used for informing the comorbidity research for the next possible pandemic outbreak.


Molecular modelling reveals eight novel druggable binding sites in SARS-CoV-2's spike protein
Uğur, İlke; Marion, Antoine (2020-12-6)
Spike glycoprotein (S), one of the signature proteins of the SARS-CoV-2, initiates the membrane fusion and virus entry to the host cell. The S protein’s key role in virus viability makes it an attractive candidate for drug design studies. Besides the recent structural characterization of the S protein, information fundamental to drug design such as possible binding sites or molecular fragments with high affinity towards the protein is unknown. We explored the druggability of this protein, focusing on its S1...
Network based approaches reveal cell subtype specific roles of astrocytes and oligodendrocytes during alzheimer’s disease
Ilgun, Atilay; Cakir, Tunahan (Orta Doğu Teknik Üniversitesi Enformatik Enstitüsü; 2022-10)
NETWORK BASED APPROACHES REVEAL CELL SUBTYPE SPECIFIC ROLES OF ASTROCYTES AND OLIGODENDROCYTES DURING ALZHEIMER’S DISEASE Atilay Ilgun1, Tunahan Cakir1 1 Department of Bioengineering, Gebze Technical University 41400, Gebze/KOCAELI e-mail: Although, Alzheimer’s disease seems to be neuron-based-disorder, contribution of glial cells to AD pathology is undeniable [1,2]. Investigation of transcriptomic changes on glial cells during AD is important since glial cells have many function...
Differential immune activation following encapsulation of immunostimulatory CpG oligodeoxynucleotide in nanoliposomes.
Erikçi, E; Gürsel, Mayda; Gürsel, I (2011-02-01)
The immunogenicity of a vaccine formulation is closely related to the effective internalization by the innate immune cells that provide prolonged and simultaneous delivery of antigen and adjuvant to relevant antigen presenting cells. Endosome associated TLR9 recognizes microbial unmethylated CpG DNA. Clinical applications of TLR9 ligands are significantly hampered due to their pre-mature in vivo digestion and rapid clearance. Liposome encapsulation is a powerful tool to increase in vivo stability as well as...
Cytotoxic T Lymphocyte Activation Signals Modulate Cytoskeletal Dynamics and Mechanical Force Generation
Pathni, Aashli; Özçelikkale, Altuğ; Rey-Suarez, Ivan; Li, Lei; Davis, Scott; Rogers, Nate; Xiao, Zhengguo; Upadhyaya, Arpita (2022-03-01)
Cytotoxic T lymphocytes (CTLs) play an integral role in the adaptive immune response by killing infected cells. Antigen presenting cells (APCs), such as dendritic cells, present pathogenic peptides to the T cell receptor on the CTL surface and co-stimulatory signals required for complete activation. Activated CTLs secrete lytic granules containing enzymes that trigger target cell death at the CTL-target contact, also known as the immune synapse (IS). The actin and microtubule cytoskeletons are instrumental ...
Molecular cloning, characterization, and homologous expression of an endochitinase gene from Bacillus thuringiensis serovar morrisoni
Okay, Sezer; Özcengiz, Gülay (2011-01-01)
The endochitinase gene (chi3023) of Bacillus thuringiensis (Bt) serovar morrisoni strain 3023 was amplified via polymerase chain reaction (PCR) and cloned in Escherichia coli. The ORF of chi3023 (GenBank Accession Number: DQ993175) consists of 2031 nucleotides encoding a 676-residue protein with a calculated molecular mass of 74.5 kDa and a pI value of 6.0. The amino acid sequence of Chi3023 was compared with previously sequenced Bt chitinases and the phylogenetic relationships among them were determined. T...
Citation Formats
E. Taylan and P. Pir, “Network Based NLP Approach to Reveal Disease Comorbidities. COVID Case Study,” Erdemli, Mersin, TÜRKİYE, 2022, p. 3041, Accessed: 00, 2023. [Online]. Available: